Eye gaze prediction for split xr

Abstract

This disclosure provides systems, devices, apparatus, and methods, including computer programs encoded on storage media, for eye gaze prediction for split XR. A processor may obtain a first indication of a first set of eye gazes associated with a user device. The processor may obtain a second indication of a second set of head poses associated with the user device. The processor may determine a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes. The processor may output a third indication of the determined third set of predicted eye gazes.

Description

EYE GAZE PREDICTION FOR SPLIT XRCROSS REFERENCE TO RELATED APPLICATION

[0001] This application claims the benefit of and priority to Indian Provisional Patent Application No. 202341060303 entitled "EYE GAZE PREDICTION FOR SPLIT XR" and filed on September 7, 2023, which is expressly incorporated by reference herein in its entirety.TECHNICAL FIELD

[0002] The present disclosure relates generally to processing systems, and more particularly, to one or more techniques for graphics processing.INTRODUCTION

[0003] Computing devices often perform graphics and / or display processing (e g., utilizing a graphics processing unit (GPU), a central processing unit (CPU), a display processor, etc.) to render and display visual content. Such computing devices may include, for example, computer workstations, mobile phones such as smartphones, embedded systems, personal computers, tablet computers, and video game consoles. GPUs are configured to execute a graphics processing pipeline that includes one or more processing stages, which operate together to execute graphics processing commands and output a frame. A central processing unit (CPU) may control the operation of the GPU by issuing one or more graphics processing commands to the GPU. Modern day CPUs are typically capable of executing multiple applications concurrently, each of which may need to utilize the GPU during execution. A display processor may be configured to convert digital information received from a CPU to analog values and may issue commands to a display panel for displaying the visual content. A device that provides content for visual presentation on a display may utilize a CPU, a GPU, and / or a display processor.

[0004] Current techniques for extended reality (XR) rendering may be based on inaccurate eye gaze predictions. There is a need for improved techniques for XR rendering.BRIEF SUMMARY

[0005] The following presents a simplified summary of one or more aspects in order to provide a basic understanding of such aspects. This summary is not an extensive overview of all contemplated aspects, and is intended to neither identify key or critical elements of all aspects nor delineate the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that is presented later.

[0006] In an aspect of the disclosure, a method, a computer-readable medium, and an apparatus are provided. The apparatus includes a memory; and a processor coupled to the memory and, based on information stored in the memory, the processor may be configured to obtain an indication of an eye gaze associated with a user device. The processor may be configured to obtain an indication of a head pose associated with the user device. The processor may be configured to determine a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose. The processor may be configured to output an indication of the determined set of predicted eye gazes.

[0007] In an aspect of the disclosure, a method, a computer-readable medium, and an apparatus are provided. The apparatus includes a memory; and a processor coupled to the memory and, based on information stored in the memory, the processor may be configured to obtain a first indication of a first set of eye gazes associated with a user device. The processor may be configured to obtain a second indication of a second set of head poses associated with the user device. The processor may be configured to determine a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes. The processor may be configured to output a third indication of the determined third set of predicted eye gazes.

[0008] In some aspects, the techniques described herein relate to a method of graphics processing, including: obtaining a first indication of a first set of eye gazes associated with a user device; obtaining a second indication of a second set of head poses associated with the user device; determining a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes; and outputting a third indication of the determined third set of predicted eye gazes.

[0009] In some aspects, the techniques described herein relate to a method, further including: correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes.

[0010] In some aspects, the techniques described herein relate to a method, where correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes includes: correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes based on a first set of time indicators associated with the first set of eye gazes and a second set of time indicators associated with the second set of head poses.

[0011] In some aspects, the techniques described herein relate to a method, where the first set of eye gazes includes eye gaze time series data, where the second set of head poses includes head pose time series data.

[0012] In some aspects, the techniques described herein relate to a method, further including: obtaining a fourth indication of a fourth set of blinks associated with the user device; and selecting a third subset of the obtained first set of eye gazes based on the fourth set of blinks associated with the user device, where determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes includes: determining the third set of predicted eye gazes based on the selected third subset of the obtained first set of eye gazes and a fourth subset of the obtained second set of head poses correlated with the selected third subset of the obtained first set of eye gazes.

[0013] In some aspects, the techniques described herein relate to a method, further including: determining that a third subset of the obtained first set of eye gazes is associated with a saccade; and selecting a fourth subset of the obtained first set of eye gazes based on the determined third subset of the obtained first set of eye gazes associated with the saccade, where determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes includes: determining the third set of predicted eye gazes based on the selected fourth subset of the obtained first set of eye gazes and a fifth subset of the obtained second set of head poses correlated with the selected fourth subset of the obtained first set of eye gazes.

[0014] In some aspects, the techniques described herein relate to a method, further including: converting a third subset of the first set of eye gazes into a fourth set of sixdimensional (6D) representations of the first set of eye gazes, where determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes includes: determining the third set of predicted eye gazes based on the converted fourth set of 6D representations of the first set of eye gazes.

[0015] In some aspects, the techniques described herein relate to a method, where determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes includes: determining the third set of predicted eye gazes based on a first set of fully-connected network (FCN) blocks and a set of gated recurrent unit (GRU) blocks.

[0016] In some aspects, the techniques described herein relate to a method, where determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes includes: determining the third set of predicted eye gazes based on a predictor model that uses the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes as inputs.

[0017] In some aspects, the techniques described herein relate to a method, where the predictor model includes at least one of a regression model, a time series forecaster model, or a neural network.

[0018] In some aspects, the techniques described herein relate to a method, further including: determining that a third subset of the first set of eye gazes is associated with a user tracking a virtual object; determining a fourth set of predicted eye gazes based on the determined third subset of the first set of eye gazes with respect to a head of the user; and outputting a fourth indication of the determined fourth set of predicted eye gazes.

[0019] In some aspects, the techniques described herein relate to a method, further including: determining that a fourth subset of the first set of eye gazes are not associated with the user tracking the virtual object, where determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtainedsecond set of head poses correlated with the first subset of the obtained first set of eye gazes includes: determining the third set of predicted eye gazes based on the determined fourth subset of the first set of eye gazes and a fifth subset of the obtained second set of head poses correlated with the determined fourth subset of the first set of eye gazes.

[0020] In some aspects, the techniques described herein relate to a method, where obtaining the first set of eye gazes includes: generating, via an eye gaze camera associated with the user device via, the first set of eye gazes at a first sampling rate, where obtaining the second set of head poses includes: generating, via an inertial measurement unit (IMU) associated with the user device, the second set of head poses at a second sampling rate.

[0021] In some aspects, the techniques described herein relate to a method, where the first sampling rate (e g., 90 Hz) is less than the second sampling rate (e g , 1000 Hz).

[0022] In some aspects, the techniques described herein relate to a method, where the first set of eye gazes includes a third set of eye gaze time series data, where the second set of head poses includes a fourth set of head pose time series data.

[0023] In some aspects, the techniques described herein relate to a method, where the first set of eye gazes includes first translation information and first orientation information of at least one eye of a user associated with the user device, where the second set of head poses includes second translation information and second orientation information of a head of the user associated with the user device.

[0024] In some aspects, the techniques described herein relate to a method, where outputting the third indication of the determined third set of predicted eye gazes includes: transmitting the third indication of the determined third set of predicted eye gazes; receiving a fourth indication of a fourth set of rendered frames based on a transmission of the third indication of the determined third set of predicted eye gazes; and outputting a fifth indication of the fourth set of rendered frames.

[0025] In some aspects, the techniques described herein relate to a method, where outputting the fifth indication of the fourth set of rendered frames includes: storing the fourth set of rendered frames in at least one of a memory, a buffer, or a cache.

[0026] In some aspects, the techniques described herein relate to a method, further including: decoding the fourth set of rendered frames; and warping the decoded fourth set of rendered frames based on a latest available head pose of a user associated with theuser device, where outputting the fifth indication of the fourth set of rendered frames includes: outputting a sixth indication of a fifth set of warped, decoded rendered frames.

[0027] In some aspects, the techniques described herein relate to a method, where each of the fourth set of rendered frames includes a first region and a second region, where the first region is associated with a first quality level and the second region is associated with a second quality level, where the first quality level is greater than the second quality level, and wherein the first region corresponds to at least one of the third set of predicted eye gazes.

[0028] In some aspects, the techniques described herein relate to a method, where obtaining the first indication of the first set of eye gazes associated with the user device includes: receiving the first indication of the first set of eye gazes from the user device, where obtaining the second indication of the second set of head poses associated with the user device includes: receiving the second indication of the second set of head poses from the user device.

[0029] In some aspects, the techniques described herein relate to a method, where outputting the third indication of the determined third set of predicted eye gazes includes: outputting the third indication to a Tenderer, further including: rendering, via the Tenderer, a fourth set of frames based on the determined third set of predicted eye gazes; and transmitting the rendered fourth set of frames to the user device.

[0030] In some aspects, the techniques described herein relate to a method, where rendering the fourth set of frames based on the determined third set of predicted eye gazes includes: rendering a first region and a second region of each of the fourth set of frames, where the first region is associated with a first quality level and the second region is associated with a second quality level, where the first quality level is greater than the second quality level, and wherein the first region corresponds with at least one of the determined third set of predicted eye gazes.

[0031] In some aspects, the techniques described herein relate to a method, where transmitting the rendered fourth set of frames to the user device includes: encoding the rendered fourth set of frames; and transmitting the encoded rendered fourth set of frames.

[0032] In an aspect of the disclosure, a method, a computer-readable medium, and an apparatus are provided. The apparatus includes a memory; and a processor coupled tothe memory and, based on information stored in the memory, the processor is configured to: provide, at a first time instance, a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user; obtain, based on the first indication and the second indication, a rendered frame that is based on a predicted eye gaze of the user at a second time instance that occurs after the first time instance; and output an indication of the rendered frame.

[0033] In an aspect of the disclosure, a method, a computer-readable medium, and an apparatus are provided. The apparatus includes a memory; and a processor coupled to the memory and, based on information stored in the memory, the processor is configured to: obtain a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user, where the first indication and the second indication correspond to a first time instance; generate, based on the first indication and the second indication, a predicted eye gaze of the user, where the predicted eye gaze of the user corresponds to a second time instance that occurs after the first time instance; render a frame based on the predicted eye gaze of the user; and output the rendered frame.

[0034] In some aspects, the techniques described herein relate to a method for graphics processing, including: obtaining an indication of an eye gaze associated with a user device; obtaining an indication of a head pose associated with the user device; determining a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose; and outputting an indication of the determined set of predicted eye gazes.

[0035] In some aspects, the techniques described herein relate to a method, further including: correlating the head pose with the eye gaze.

[0036] In some aspects, the techniques described herein relate to a method, wherein correlating the head pose with the eye gaze includes: correlating the head pose with the eye gaze based on a first time indicator associated with the eye gaze and a second time indicator associated with the head pose.

[0037] In some aspects, the techniques described herein relate to a method, wherein the indication of the eye gaze includes eye gaze time series data, wherein the indication of the head pose includes head pose time series data.

[0038] In some aspects, the techniques described herein relate to a method, further including: obtaining an indication of a second eye gaze associated with the user device; obtainingan indication of a set of blinks associated with the user device; and determining that the second eye gaze is associated with at least one of the set of blinks, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose includes: refraining from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of blinks.

[0039] In some aspects, the techniques described herein relate to a method, further including: obtaining an indication of a second eye gaze associated with the user device; obtaining an indication of a set of saccades associated with the user device; and determining that the second eye gaze is associated with at least one of the set of saccades, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose includes: refraining from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of saccades.

[0040] In some aspects, the techniques described herein relate to a method, further including: converting the eye gaze into a six dimensional (6D) representation of the eye gaze, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose includes: determining the set of predicted eye gazes based on the converted 6D representation of the eye gaze.

[0041] In some aspects, the techniques described herein relate to a method, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose includes: determining the set of predicted eye gazes based on a first set of fully-connected network (FCN) blocks and a set of gated recurrent unit (GRU) blocks.

[0042] In some aspects, the techniques described herein relate to a method, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose includes: determining the set of predicted eye gazes based on a predictor model that uses the eye gaze and the head pose as inputs.

[0043] In some aspects, the techniques described herein relate to a method, wherein the predictor model includes at least one of a regression model, a time series forecaster model, or a neural network.

[0044] In some aspects, the techniques described herein relate to a method, further including: obtaining an indication of a second eye gaze associated with the user device; determining that the second eye gaze is associated with a user tracking a virtual obj ect; determining a second set of predicted eye gazes based on the obtained indication of the second eye gaze with respect to a head of the user in response to a determination that the eye gaze is associated with the user tracking the virtual object; and output an indication of the determined second set of predicted eye gazes.

[0045] In some aspects, the techniques described herein relate to a method, further including: determining that the eye gaze is not associated with the user tracking the virtual object, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose includes: determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose includes in response to a determination that the eye gaze is not associated with the user tracking the virtual object.

[0046] In some aspects, the techniques described herein relate to a method, wherein obtaining the indication of the eye gaze associate with the user device includes: generating, via an eye gaze camera associated with the user device, a set of eye gazes including the eye gaze at a first sampling rate, wherein obtaining the head pose includes: generating, via an inertial measurement unit (IMU) associated with the user device, a set of head poses including the head pose at a second sampling rate.

[0047] In some aspects, the techniques described herein relate to a method, wherein the first sampling rate is less than the second sampling rate.

[0048] In some aspects, the techniques described herein relate to a method, wherein the indication of the eye gaze includes first translation information and first orientation information of at least one eye of a user associated with the user device, wherein the indication of the head pose includes second translation information and second orientation information of a head of the user associated with the user device.

[0049] In some aspects, the techniques described herein relate to a method, wherein outputting the indication of the determined set of predicted eye gazes includes: transmitting the indication of the determined set of predicted eye gazes; receiving anindication of a set of rendered frames after a transmission of the indication of the determined set of predicted eye gazes; and outputting an indication of the set of rendered frames.

[0050] In some aspects, the techniques described herein relate to a method, wherein outputting the indication of the set of rendered frames includes: storing the set of rendered frames in at least one of a memory, a buffer, or a cache.

[0051] In some aspects, the techniques described herein relate to a method, further including: decoding the set of rendered frames; and warping the decoded set of rendered frames based on a latest available head pose of a user associated with the user device, wherein outputting the indication of the set of rendered frames includes: outputting an indication of a set of warped, decoded rendered frames.

[0052] In some aspects, the techniques described herein relate to a method, wherein each of the set of rendered frames includes a first region and a second region, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds to at least one of the set of predicted eye gazes.

[0053] In some aspects, the techniques described herein relate to a method, wherein obtaining the indication of the eye gaze associated with the user device includes: receiving the indication of the eye gaze from the user device, wherein obtaining the indication of the head pose associated with the user device includes: receiving the indication of the head pose from the user device.

[0054] In some aspects, the techniques described herein relate to a method, wherein outputting the indication of the determined set of predicted eye gazes includes: outputting the indication to a Tenderer, wherein the method further includes: rendering, via the Tenderer, a set of frames based on the determined set of predicted eye gazes; and transmitting the rendered set of frames to the user device.

[0055] In some aspects, the techniques described herein relate to a method, wherein rendering the set of frames based on the determined set of predicted eye gazes includes: rendering a first region and a second region of each of the set of frames, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the secondquality level, and wherein the first region corresponds with at least one of the determined set of predicted eye gazes.

[0056] In some aspects, the techniques described herein relate to a method, wherein transmitting the rendered set of frames to the user device includes: encoding the rendered set of frames; and transmitting the encoded rendered set of frames.

[0057] To the accomplishment of the foregoing and related ends, the one or more aspects include the features hereinafter fully described and particularly pointed out in the claims. The following description and the annexed drawings set forth in detail certain illustrative features of the one or more aspects. These features are indicative, however, of but a few of the various ways in which the principles of various aspects may be employed, and this description is intended to include all such aspects and their equivalents.BRIEF DESCRIPTION OF THE DRAWINGS

[0058] FIG. 1 is a block diagram that illustrates an example content generation system in accordance with one or more techniques of this disclosure.

[0059] FIG. 2 illustrates an example graphics processor (e.g., a graphics processing unit (GPU)) in accordance with one or more techniques of this disclosure.

[0060] FIG. 3 illustrates an example image or surface in accordance with one or more techniques of this disclosure.

[0061] FIG. 4 is a diagram illustrating an example of a split extended reality (XR) system in accordance with one or more techniques of this disclosure.

[0062] FIG. 5 is a diagram illustrating example aspects pertaining to a prediction model for eye gaze prediction in accordance with one or more techniques of this disclosure.

[0063] FIG. 6 is a diagram illustrating example simulation results pertaining to a prediction model in accordance with one or more techniques of this disclosure.

[0064] FIG. 7A is a diagram illustrating example aspects pertaining to using XR metadata to enhance an eye gaze prediction in accordance with one or more techniques of this disclosure.

[0065] FIG. 7B is a diagram illustrating example aspects pertaining to using XR metadata to enhance an eye gaze prediction in accordance with one or more techniques of this disclosure.

[0066] FIG. 8 is a flowchart illustrating an example workflow to predict a gaze of a user, in accordance with one or more techniques of this disclosure.

[0067] FIG. 9 is a diagram illustrating an example aspect of a set of fully connected network (FCN) blocks and a set of gated recurrent unit (GRU) blocks configured to predict a gaze of a user, in accordance with one or more techniques of this disclosure.

[0068] FIG. 10 is a diagram illustrating an example aspect of a gaze predictor design, in accordance with one or more techniques of this disclosure.

[0069] FIG. 11 is a connection flow diagram illustrating an example of a server configured to provide a predicted gaze to a client, in accordance with one or more techniques of this disclosure.

[0070] FIG. 12 is a connection flow diagram illustrating an example of a server configured to provide a predicted gaze to a client, in accordance with one or more techniques of this disclosure.

[0071] FIG. 13 is a flowchart of an example method of graphics processing in accordance with one or more techniques of this disclosure.

[0072] FIG. 14 is a flowchart of an example method of graphics processing in accordance with one or more techniques of this disclosure.

[0073] FIG. 15 is a flowchart of an example method of graphics processing in accordance with one or more techniques of this disclosure.DETAILED DESCRIPTION

[0074] Various aspects of systems, apparatuses, computer program products, and methods are described more fully hereinafter with reference to the accompanying drawings. This disclosure may, however, be embodied in many different forms and should not be construed as limited to any specific structure or function presented throughout this disclosure. Rather, these aspects are provided so that this disclosure will be thorough and complete, and will fully convey the scope of this disclosure to those skilled in the art. Based on the teachings herein one skilled in the art should appreciate that the scope of this disclosure is intended to cover any aspect of the systems, apparatuses, computer program products, and methods disclosed herein, whether implemented independently of, or combined with, other aspects of the disclosure. For example, an apparatus may be implemented or a method may be practiced using any number of the aspects set forth herein. In addition, the scope of the disclosure is intended to coversuch an apparatus or method which is practiced using other structure, functionality, or structure and functionality in addition to or other than the various aspects of the disclosure set forth herein. Any aspect disclosed herein may be embodied by one or more elements of a claim.

[0075] Although various aspects are described herein, many variations and permutations of these aspects fall within the scope of this disclosure. Although some potential benefits and advantages of aspects of this disclosure are mentioned, the scope of this disclosure is not intended to be limited to particular benefits, uses, or objectives. Rather, aspects of this disclosure are intended to be broadly applicable to different wireless technologies, system configurations, processing systems, networks, and transmission protocols, some of which are illustrated by way of example in the figures and in the following description. The detailed description and drawings are merely illustrative of this disclosure rather than limiting, the scope of this disclosure being defined by the appended claims and equivalents thereof.

[0076] Several aspects are presented with reference to various apparatus and methods. These apparatus and methods are described in the following detailed description and illustrated in the accompanying drawings by various blocks, components, circuits, processes, algorithms, and the like (collectively referred to as “elements”). These elements may be implemented using electronic hardware, computer software, or any combination thereof. Whether such elements are implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system.

[0077] By way of example, an element, or any portion of an element, or any combination of elements may be implemented as a “processing system” that includes one or more processors (which may also be referred to as processing units). Examples of processors include microprocessors, microcontrollers, graphics processing units (GPUs), general purpose GPUs (GPGPUs), central processing units (CPUs), application processors, digital signal processors (DSPs), reduced instruction set computing (RISC) processors, systems-on-chip (SOCs), baseband processors, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), programmable logic devices (PLDs), state machines, gated logic, discrete hardware circuits, and other suitable hardware configured to perform the various functionality described throughout this disclosure. One or more processors in theprocessing system may execute software. Software can be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software components, applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures, functions, etc., whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise.

[0078] The term application may refer to software. As described herein, one or more techniques may refer to an application (e.g., software) being configured to perform one or more functions. In such examples, the application may be stored in a memory (e.g., on-chip memory of a processor, system memory, or any other memory). Hardware described herein, such as a processor may be configured to execute the application. For example, the application may be described as including code that, when executed by the hardware, causes the hardware to perform one or more techniques described herein. As an example, the hardware may access the code from a memory and execute the code accessed from the memory to perform one or more techniques described herein. In some examples, components are identified in this disclosure. In such examples, the components may be hardware, software, or a combination thereof. The components may be separate components or subcomponents of a single component.

[0079] In one or more examples described herein, the functions described may be implemented in hardware, software, or any combination thereof. If implemented in software, the functions may be stored on or encoded as one or more instructions or code on a computer-readable medium. Computer-readable media includes computer storage media. Storage media may be any available media that can be accessed by a computer. By way of example, and not limitation, such computer-readable media can include a random access memory (RAM), a read-only memory (ROM), an electrically erasable programmable ROM (EEPROM), optical disk storage, magnetic disk storage, other magnetic storage devices, combinations of the aforementioned types of computer-readable media, or any other medium that can be used to store computer executable code in the form of instructions or data structures that can be accessed by a computer.

[0080] As used herein, instances of the term “content” may refer to “graphical content,” an “image,” etc., regardless of whether the terms are used as an adjective, noun, or otherparts of speech. In some examples, the term “graphical content,” as used herein, may refer to a content produced by one or more processes of a graphics processing pipeline. In further examples, the term “graphical content,” as used herein, may refer to a content produced by a processing unit configured to perform graphics processing. In still further examples, as used herein, the term “graphical content” may refer to a content produced by a graphics processing unit.

[0081] Split rendering (i.e., a split rendering process) may refer to a paradigm whereby a client device (e.g., a head mounted unit (HMU), head mounted display (HMD)) and a server collaborate to facilitate the display of graphical content on a display of the client device, where the client device and the server may be in wired communication and / or in wireless communication with one another. For instance, a portion of rendering tasks (or other tasks) may be offloaded to the server. The server may perform the rendering tasks (or the other tasks). The server may transmit an output of the rendering tasks (or the other tasks) to the client device. The client device may perform additional processing based on the received output in order to display the graphical content. Split rendering (e.g., split XR rendering) may enable the client device to present relatively high quality graphical content on a display while conserving battery life of the client device.

[0082] In a split rendering system, a client device may generate an eye gaze of an eye of a user of the client device and transmit the eye gaze to a server. The server may use the eye gaze to render and encode a frame. For instance, the server may render and encode the frame at a higher quality in a foveal region of the frame and the server may render and encode the frame at a lower quality in a non-foveal region of the frame. However, the eye gaze of the user may change between a time at which the eye gaze is generated and a time at which the rendered frame is displayed. As a result, the user may focus on a non-foveal region of the frame (e.g., a lower quality region of the frame), which may affect user experience.

[0083] Various technologies pertaining to eye gaze prediction for split extended reality (XR) are described herein. In an example, an apparatus (e.g., a client device) provides, at a first time instance, a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user. The apparatus (e.g., the client device) obtains, based on the first indication and the second indication, a rendered frame that is based on a predicted eye gaze of the user at a second timeinstance that occurs after the first time instance. The apparatus (e g., the client device) outputs an indication of the rendered frame. Vis-a-vis providing the first indication of the set of head poses and the second indication of the set of eye gazes, the apparatus (e.g., the client device) may obtain a rendered frame that is rendered based on a predicted eye gaze of the user at a time at which the frame is to be displayed, which may improve user experience.

[0084] In another example, an apparatus (e g., a server) obtains a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user, where the first indication and the second indication correspond to a first time instance. The apparatus (e g., the server) generates, based on the first indication and the second indication, a predicted eye gaze of the user, where the predicted eye gaze of the user corresponds to a second time instance that occurs after the first time instance The apparatus (e g , the server) renders a frame based on the predicted eye gaze of the user. The apparatus (e g., the server) outputs the rendered frame. Vis-a-vis generating the predicted eye gaze of the user based on the first indication of the set of head poses and the second indication of the set of eye gazes, the apparatus (e.g., the server) may generate a more accurate predicted eye gaze in comparison to an apparatus that predicts an eye gaze based on eye gazes (without head poses). Furthermore, as the frame is rendered based on the predicted eye gaze of the user, a user experience may be improved.

[0085] In another example, an apparatus may be configured to obtain an indication of an eye gaze associated with a user device. The apparatus may be configured to obtain an indication of a head pose associated with the user device. The apparatus may be configured to determine a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose. The apparatus may be configured to output an indication of the determined set of predicted eye gazes.

[0086] In some aspects, an eye gaze predictor may leverage a head pose prediction to aid in an eye gaze prediction in a split-XR system. Since head pose data may be available at a higher sampling rate than eye gaze data, an eye gaze predictor may use the high- sampling rate head pose data to predict an eye gaze. An auto regressive model may be used on past eye gazes (e.g., past 3 frames) to predict future cases (e.g., next 3 frames). The head pose rotation may additionally be used as an input to the auto regressive model. XR metadata may also be used as an input to the auto regressivemodel to enhance the prediction, such as depth of objects and motion vectors from the encoder.

[0087] The examples describe herein may refer to a use and functionality of a graphics processing unit (GPU). As used herein, a GPU can be any type of graphics processor, and a graphics processor can be any type of processor that is designed or configured to process graphics content. For example, a graphics processor or GPU can be a specialized electronic circuit that is designed for processing graphics content. As an additional example, a graphics processor or GPU can be a general purpose processor that is configured to process graphics content.

[0088] FIG. 1 is a block diagram that illustrates an example content generation system 100 configured to implement one or more techniques of this disclosure. The content generation system 100 includes a device 104. The device 104 may include one or more components or circuits for performing various functions described herein. In some examples, one or more components of the device 104 may be components of a SOC. The device 104 may include one or more components configured to perform one or more techniques of this disclosure. In the example shown, the device 104 may include a processing unit 120, a content encoder / decoder 122, and a system memory 124. In some aspects, the device 104 may include a number of components (e.g., a communication interface 126, a transceiver 132 or an antenna, a receiver 128, a transmitter 130, a display processor 127, and one or more displays 131). Display(s) 131 may refer to one or more displays 131. For example, the display 131 may include a single display or multiple displays, which may include a first display and a second display. The first display may be a left-eye display and the second display may be a right-eye display. In some examples, the first display and the second display may receive different frames for presentment thereon. In other examples, the first and second display may receive the same frames for presentment thereon. In further examples, the results of the graphics processing may not be displayed on the device, e.g., the first display and the second display may not receive any frames for presentment thereon. Instead, the frames or graphics processing results may be transferred to another device. In some aspects, this may be referred to as splitrendering.

[0089] The processing unit 120 may include an internal memory 121. The processing unit 120 may be configured to perform graphics processing using a graphics processingpipeline 107. The content encoder / decoder 122 may include an internal memory 123. In some examples, the device 104 may include a processor, which may be configured to perform one or more display processing techniques on one or more frames generated by the processing unit 120 before the frames are displayed by the one or more displays 131. While the processor in the example content generation system 100 is configured as a display processor 127, it should be understood that the display processor 127 is one example of the processor and that other types of processors, controllers, etc., may be used as substitute for the display processor 127. The display processor 127 may be configured to perform display processing. For example, the display processor 127 may be configured to perform one or more display processing techniques on one or more frames generated by the processing unit 120. The one or more displays 131 may be configured to display or otherwise present frames processed by the display processor 127. In some examples, the one or more displays 131 may include one or more of a liquid crystal display (LCD), a plasma display, an organic light emitting diode (OLED) display, a projection display device, an augmented reality display device, a virtual reality display device, a head-mounted display, or any other type of display device.

[0090] Memory external to the processing unit 120 and the content encoder / decoder 122, such as system memory 124, may be accessible to the processing unit 120 and the content encoder / decoder 122. For example, the processing unit 120 and the content encoder / decoder 122 may be configured to read from and / or write to external memory, such as the system memory 124. The processing unit 120 may be communicatively coupled to the system memory 124 over a bus. In some examples, the processing unit 120 and the content encoder / decoder 122 may be communicatively coupled to the internal memory 121 over the bus or via a different connection.

[0091] The content encoder / decoder 122 may be configured to receive graphical content from any source, such as the system memory 124 and / or the communication interface 126. The system memory 124 may be configured to store received encoded or decoded graphical content. The content encoder / decoder 122 may be configured to receive encoded or decoded graphical content, e.g., from the system memory 124 and / or the communication interface 126, in the form of encoded pixel data. The content encoder / decoder 122 may be configured to encode or decode any graphical content.

[0092] The internal memory 121 or the system memory 124 may include one or more volatile or non-volatile memories or storage devices. In some examples, internal memory 121 or the system memory 124 may include RAM, static random access memory (SRAM), dynamic random access memory (DRAM), erasable programmable ROM (EPROM), EEPROM, flash memory, a magnetic data media or an optical storage media, or any other type of memory. The internal memory 121 or the system memory 124 may be a non-transitory storage medium according to some examples. The term “non- transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. However, the term “non-transitory” should not be interpreted to mean that internal memory 121 or the system memory 124 is non-movable or that its contents are static. As one example, the system memory 124 may be removed from the device 104 and moved to another device. As another example, the system memory 124 may not be removable from the device 104.

[0093] The processing unit 120 may be a CPU, a GPU, a GPGPU, or any other processing unit that may be configured to perform graphics processing. In some examples, the processing unit 120 may be integrated into a motherboard of the device 104. In further examples, the processing unit 120 may be present on a graphics card that is installed in a port of the motherboard of the device 104, or may be otherwise incorporated within a peripheral device configured to interoperate with the device 104. The processing unit 120 may include one or more processors, such as one or more microprocessors, GPUs, ASICs, FPGAs, arithmetic logic units (ALUs), DSPs, discrete logic, software, hardware, firmware, other equivalent integrated or discrete logic circuitry, or any combinations thereof. If the techniques are implemented partially in software, the processing unit 120 may store instructions for the software in a suitable, non-transitory computer-readable storage medium, e.g., internal memory 121, and may execute the instructions in hardware using one or more processors to perform the techniques of this disclosure. Any of the foregoing, including hardware, software, a combination of hardware and software, etc., may be considered to be one or more processors.

[0094] The content encoder / decoder 122 may be any processing unit configured to perform content decoding. In some examples, the content encoder / decoder 122 may be integrated into a motherboard of the device 104. The content encoder / decoder 122 may include one or more processors, such as one or more microprocessors, applicationspecific integrated circuits (ASICs), field programmable gate arrays (FPGAs), arithmetic logic units (ALUs), digital signal processors (DSPs), video processors, discrete logic, software, hardware, firmware, other equivalent integrated or discrete logic circuitry, or any combinations thereof. If the techniques are implemented partially in software, the content encoder / decoder 122 may store instructions for the software in a suitable, non-transitory computer-readable storage medium, e.g., internal memory 123, and may execute the instructions in hardware using one or more processors to perform the techniques of this disclosure. Any of the foregoing, including hardware, software, a combination of hardware and software, etc., may be considered to be one or more processors.

[0095] In some aspects, the content generation system 100 may include a communication interface 126. The communication interface 126 may include a receiver 128 and a transmitter 130. The receiver 128 may be configured to perform any receiving function described herein with respect to the device 104. Additionally, the receiver 128 may be configured to receive information, e.g., eye or head position information, rendering commands, and / or location information, from another device. The transmitter 130 may be configured to perform any transmitting function described herein with respect to the device 104. For example, the transmitter 130 may be configured to transmit information to another device, which may include a request for content. The receiver 128 and the transmitter 130 may be combined into a transceiver 132. In such examples, the transceiver 132 may be configured to perform any receiving function and / or transmitting function described herein with respect to the device 104.

[0096] Referring again to FIG. 1, in certain aspects, the processing unit 120 may include an eye gaze predictor 198 configured to provide, at a first time instance, a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user; obtain, based on the first indication and the second indication, a rendered frame that is based on a predicted eye gaze of the user at a second time instance that occurs after the first time instance; and output an indication of the rendered frame. In certain aspects, the processing unit 120 may include an eye gaze predictor 199 configured to obtain a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user, where the first indication and the second indication correspond to a first time instance; generate,based on the first indication and the second indication, a predicted eye gaze of the user, where the predicted eye gaze of the user corresponds to a second time instance that occurs after the first time instance; render a frame based on the predicted eye gaze of the user; and output the rendered frame.

[0097] In other aspects, the eye gaze predictor 198 may be configured to obtain an indication of an eye gaze associated with a user device. The eye gaze predictor 198 may be configured to obtain an indication of a head pose associated with the user device. The eye gaze predictor 198 may be configured to determine a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose. The eye gaze predictor 198 may be configured to output an indication of the determined set of predicted eye gazes.

[0098] Although the following description may be focused on graphics processing, the concepts described herein may be applicable to other similar processing techniques. Furthermore, although the following description may be focused on split rendering, the concepts described herein may also be applicable to non-split rendering.

[0099] A device, such as the device 104, may refer to any device, apparatus, or system configured to perform one or more techniques described herein. For example, a device may be a server, a base station, a user equipment, a client device, a station, an access point, a computer such as a personal computer, a desktop computer, a laptop computer, a tablet computer, a computer workstation, or a mainframe computer, an end product, an apparatus, a phone, a smart phone, a server, a video game platform or console, a handheld device such as a portable video game device or a personal digital assistant (PDA), a wearable computing device such as a smart watch, an augmented reality device, or a virtual reality device, a non-wearable device, a display or display device, a television, a television set-top box, an intermediate network device, a digital media player, a video streaming device, a content streaming device, an in-vehicle computer, any mobile device, any device configured to generate graphical content, or any device configured to perform one or more techniques described herein. Processes herein may be described as performed by a particular component (e g., a GPU) but in other embodiments, may be performed using other components (e g., a CPU) consistent with the disclosed embodiments.

[0100] GPUs can process multiple types of data or data packets in a GPU pipeline. For instance, in some aspects, a GPU can process two types of data or data packets, e.g.,context register packets and draw call data. A context register packet can be a set of global state information, e.g., information regarding a global register, shading program, or constant data, which can regulate how a graphics context will be processed. For example, context register packets can include information regarding a color format. In some aspects of context register packets, there can be a bit or bits that indicate which workload belongs to a context register. Also, there can be multiple functions or programming running at the same time and / or in parallel. For example, functions or programming can describe a certain operation, e.g., the color mode or color format. Accordingly, a context register can define multiple states of a GPU.

[0101] Context states can be utilized to determine how an individual processing unit functions, e.g., a vertex fetcher (VFD), a vertex shader (VS), a shader processor, or a geometry processor, and / or in what mode the processing unit functions. In order to do so, GPUs can use context registers and programming data. In some aspects, a GPU can generate a workload, e.g., a vertex or pixel workload, in the pipeline based on the context register definition of a mode or state. Certain processing units, e.g., a VFD, can use these states to determine certain functions, e.g., how a vertex is assembled. As these modes or states can change, GPUs may need to change the corresponding context. Additionally, the workload that corresponds to the mode or state may follow the changing mode or state.

[0102] FIG. 2 illustrates an example GPU 200 in accordance with one or more techniques of this disclosure. As shown in FIG. 2, GPU 200 includes command processor (CP) 210, draw call packets 212, VFD 220, VS 222, vertex cache (VPC) 224, triangle setup engine (TSE) 226, rasterizer (RAS) 228, Z process engine (ZPE) 230, pixel interpolator (PI) 232, fragment shader (FS) 234, render backend (RB) 236, L2 cache (UCHE) 238, and system memory 240. Although FIG. 2 displays that GPU 200 includes processing units 220-238, GPU 200 can include a number of additional processing units. Additionally, processing units 220-238 are merely an example and any combination or order of processing units can be used by GPUs according to the present disclosure. GPU 200 also includes command buffer 250, context register packets 260, and context states 261.

[0103] As shown in FIG. 2, a GPU can utilize a CP, e.g., CP 210, or hardware accelerator to parse a command buffer into context register packets, e.g., context register packets 260, and / or draw call data packets, e.g., draw call packets 212. The CP 210 can thensend the context register packets 260 or draw call packets 212 through separate paths to the processing units or blocks in the GPU. Further, the command buffer 250 can alternate different states of context registers and draw calls. For example, a command buffer can simultaneously store the following information: context register of context N, draw call(s) of context N, context register of context N+l, and draw call(s) of context N+l.

[0104] GPUs can render images in a variety of different ways. In some instances, GPUs can render an image using direct rendering and / or tiled rendering. In tiled rendering GPUs, an image can be divided or separated into different sections or tiles. After the division of the image, each section or tile can be rendered separately. Tiled rendering GPUs can divide computer graphics images into a grid format, such that each portion of the grid, i.e., a tile, is separately rendered. In some aspects of tiled rendering, during a binning pass, an image can be divided into different bins or tiles. In some aspects, during the binning pass, a visibility stream can be constructed where visible primitives or draw calls can be identified. A rendering pass may be performed after the binning pass. In contrast to tiled rendering, direct rendering does not divide the frame into smaller bins or tiles. Rather, in direct rendering, the entire frame is rendered at a single time (i.e., without a binning pass). Additionally, some types of GPUs can allow for both tiled rendering and direct rendering (e.g., flex rendering).

[0105] In some aspects, GPUs can apply the drawing or rendering process to different bins or tiles. For instance, a GPU can render to one bin, and perform all the draws for the primitives or pixels in the bin. During the process of rendering to a bin, the render targets can be located in GPU internal memory (GMEM). In some instances, after rendering to one bin, the content of the render targets can be moved to a system memory and the GMEM can be freed for rendering the next bin. Additionally, a GPU can render to another bin, and perform the draws for the primitives or pixels in that bin. Therefore, in some aspects, there might be a small number of bins, e.g., four bins, that cover all of the draws in one surface. Further, GPUs can cycle through all of the draws in one bin, but perform the draws for the draw calls that are visible, i.e., draw calls that include visible geometry. In some aspects, a visibility stream can be generated, e.g., in a binning pass, to determine the visibility information of each primitive in an image or scene. For instance, this visibility stream can identify whether a certain primitive is visible or not. In some aspects, this information can be used toremove primitives that are not visible so that the non-visible primitives are not rendered, e.g., in the rendering pass. Also, at least some of the primitives that are identified as visible can be rendered in the rendering pass.

[0106] In some aspects of tiled rendering, there can be multiple processing phases or passes. For instance, the rendering can be performed in two passes, e.g., a binning, a visibility or bin-visibility pass and a rendering or bin-rendering pass. During a visibility pass, a GPU can input a rendering workload, record the positions of the primitives or triangles, and then determine which primitives or triangles fall into which bin or area. In some aspects of a visibility pass, GPUs can also identify or mark the visibility of each primitive or triangle in a visibility stream. During a rendering pass, a GPU can input the visibility stream and process one bin or area at a time. In some aspects, the visibility stream can be analyzed to determine which primitives, or vertices of primitives, are visible or not visible. As such, the primitives, or vertices of primitives, that are visible may be processed. By doing so, GPUs can reduce the unnecessary workload of processing or rendering primitives or triangles that are not visible.

[0107] In some aspects, during a visibility pass, certain types of primitive geometry, e.g., position-only geometry, may be processed. Additionally, depending on the position or location of the primitives or triangles, the primitives may be sorted into different bins or areas. In some instances, sorting primitives or triangles into different bins may be performed by determining visibility information for these primitives or triangles. For example, GPUs may determine or write visibility information of each primitive in each bin or area, e.g., in a system memory. This visibility information can be used to determine or generate a visibility stream. In a rendering pass, the primitives in each bin can be rendered separately. In these instances, the visibility stream can be fetched from memory and used to remove primitives which are not visible for that bin.

[0108] Some aspects of GPUs or GPU architectures can provide a number of different options for rendering, e.g., software rendering and hardware rendering. In software rendering, a driver or CPU can replicate an entire frame geometry by processing each view one time. Additionally, some different states may be changed depending on the view. As such, in software rendering, the software can replicate the entire workload by changing some states that may be utilized to render for each viewpoint in an image. In certain aspects, as GPUs may be submitting the same workload multiple times for each viewpoint in an image, there may be an increased amount of overhead. Inhardware rendering, the hardware or GPU may be responsible for replicating or processing the geometry for each viewpoint in an image. Accordingly, the hardware can manage the replication or processing of the primitives or triangles for each viewpoint in an image.

[0109] FIG. 3 illustrates image or surface 300, including multiple primitives divided into multiple bins in accordance with one or more techniques of this disclosure. As shown in FIG. 3, image or surface 300 includes area 302, which includes primitives 321, 322, 323, and 324. The primitives 321, 322, 323, and 324 are divided or placed into different bins, e.g., bins 310, 311, 312, 313, 314, and 315. FIG. 3 illustrates an example of tiled rendering using multiple viewpoints for the primitives 321-324. For instance, primitives 321-324 are in first viewpoint 350 and second viewpoint 351. As such, the GPU processing or rendering the image or surface 300 including area 302 can utilize multiple viewpoints or multi-view rendering.

[0110] As indicated herein, GPUs or graphics processors can use a tiled rendering architecture to reduce power consumption or save memory bandwidth. As further stated above, this rendering method can divide the scene into multiple bins, as well as include a visibility pass that identifies the triangles that are visible in each bin. Thus, in tiled rendering, a full screen can be divided into multiple bins or tiles. The scene can then be rendered multiple times, e.g., one or more times for each bin.[oni] In aspects of graphics rendering, some graphics applications may render to a single target, i.e., a render target, one or more times. For instance, in graphics rendering, a frame buffer on a system memory may be updated multiple times. The frame buffer can be a portion of memory or random access memory (RAM), e.g., containing a bitmap or storage, to help store display data for a GPU. The frame buffer can also be a memory buffer containing a complete frame of data. Additionally, the frame buffer can be a logic buffer. In some aspects, updating the frame buffer can be performed in bin or tile rendering, where, as discussed above, a surface is divided into multiple bins or tiles and then each bin or tile can be separately rendered. Further, in tiled rendering, the frame buffer can be partitioned into multiple bins or tiles.

[0112] As indicated herein, in some aspects, such as in bin or tiled rendering architecture, frame buffers can have data stored or written to them repeatedly, e.g., when rendering from different types of memory. This can be referred to as resolving and unresolving the frame buffer or system memory. For example, when storing or writing to oneframe buffer and then switching to another frame buffer, the data or information on the frame buffer can be resolved from the GMEM at the GPU to the system memory, i.e., memory in the double data rate (DDR) RAM or dynamic RAM (DRAM).

[0113] In some aspects, the system memory can also be system-on-chip (SoC) memory or another chip-based memory to store data or information, e.g., on a device or smart phone. The system memory can also be physical data storage that is shared by the CPU and / or the GPU. In some aspects, the system memory can be a DRAM chip, e.g., on a device or smart phone. Accordingly, SoC memory can be a chip-based manner in which to store data.

[0114] In some aspects, the GMEM can be on-chip memory at the GPU, which can be implemented by static RAM (SRAM). Additionally, GMEM can be stored on a device, e.g., a smart phone. As indicated herein, data or information can be transferred between the system memory or DRAM and the GMEM, e g., at a device In some aspects, the system memory or DRAM can be at the CPU or GPU. Additionally, data can be stored at the DDR or DRAM. In some aspects, such as in bin or tiled rendering, a small portion of the memory can be stored at the GPU, e.g., at the GMEM. In some instances, storing data at the GMEM may utilize a larger processing workload and / or consume more power compared to storing data at the frame buffer or system memory.

[0115] A user may wear a display device in order to experienced extended reality (XR) content. XR may refer to a technology that blends aspects of a digital experience and the real world (i.e., XR content). XR may include augmented reality (AR), mixed reality (MR), and / or virtual reality (VR). In AR, AR objects may be superimposed on a real-world environment as perceived through the display device. In an example, AR content may be experienced through AR glasses that include a transparent or semitransparent surface. An AR object may be projected onto the transparent or semitransparent surface of the glasses as a user views an environment through the glasses. In general, the AR object may not be present in the real world and the user may not interact with the AR object. In MR, MR objects may be superimposed on a real -world environment as perceived through the display device and the user may interact with the MR objects. In some aspects, MR objects may include “video see through” with virtual content added. In an example, the user may “touch” a MR object being displayed to the user (i.e., the user may place a hand at a location in the real world where the MR object appears to be located from the perspective of the user), and theMR object may “move” based on the MR object being touched (i.e., a location of the MR object on a display may change). In general, MR content may be experienced through MR glasses (similar to AR glasses) worn by the user or through a head mounted display (HMD) worn by the user. The HMD may include a camera and one or more display panels. The HMD may capture an image of environment as perceived through the camera and display the image of the environment to the user with MR objects overlaid thereon. Unlike the transparent or semi-transparent surface of the AR / MR glasses, the one or more display panels of the HMD may not be transparent or semi-transparent. In VR, a user may experience a fully-immersive digital environment in which the real-world is blocked out. VR content may be experienced through an HMD. An XR device may refer to a device that is capable of presenting XR content.

[0116] FIG. 4 is a diagram 400 illustrating an example of a split extended reality (XR) system in accordance with one or more techniques of this disclosure. In a split XR system, a client 404, (e.g., an HMD) may transmit head mounted display (HMD) pose data (i.e., orientation information, eye gaze data, etc.) to a server 402 as the data 406. The server 402 may render XR content based on the HMD pose data and the eye gaze data, compress the XR content, and transmit the XR content to the client 404 along with a render pose as the encoded bitstream 408. The client 404 may decompress and render the received content after warping the content using the render pose and latest display pose information. The latency of the round-trip between the client 404 and the server 402 may be represented by the round-trip latency period 410.

[0117] In a split XR system, eye gazes may be generated at a client HMD at time 7’ milliseconds (ms) and the eye gazes may be sent to a server. The server may use the head pose and eye gaze to render and encode frames. The server may transmit the frame back to the client HMD, where the client HMD may decode and display a frame at time 7+7” milliseconds (ms). The server may use the eye gaze to render / encode frames at a high quality in a foveal region associated with the frame and at a lower quality outside of the foveal region; however, in either case, an eye of a user may be moving continuously. If the eye at time 7+7” has moved significantly from where the eye was at time “t,” then the user may focus the eye on a part of a frame which may have lower quality, thus deteriorating a user experience. Aspects presented herein pertain to an algorithm that may predict an eye gaze at time 7+7” accurately.

[0118] FIG. 5 is a diagram 500 illustrating example aspects pertaining to a prediction model for eye gaze prediction in accordance with one or more techniques of this disclosure. Aspects presented herein may exploit the fact that head poses and eye gazes may be generated together. Aspects presented herein may leverage head pose prediction to aid in eye gaze prediction. In an example, a head pose may be available at a high sampling rate (e.g., an inertial measurement unit (IMU) may sample at 1000 hertz (Hz)), whereas an eye gaze may be available at an eye tracking camera frames per second (fps) (e.g., 30 Hz for instance, or 33 ms apart).

[0119] In one aspect, the predictor model may be a time series forecaster. In another aspect, the predictor model may be an auto regression model. In another aspect, the predictor model may be a neural network. In another aspect, the predictor model may be a transformer.

[0120] Aspects presented herein may be associated with advantages over predicting eye gazes using a history of eye gaze samples.

[0121] In one case, a user may gaze at a stationary object in the virtual world while moving his / her head. In such a case, the eye gaze of the user may move continuously in an opposite direction of the head. Using the aspects described herein, head pose information may convert eye gazes at different time instances (eye gazes corresponding to stationary object projected onto the image) into world coordinates. The world coordinates may lie in the same location (as the object is not moving). Using this information, along with the predicted head pose at time ‘T+t,1aspects presented herein may determine an eye gaze at time ‘T+

[0122] In another case, a user may be gazing at a moving object in the virtual world and the user may move his / her head in order to track the moving object. In such a case, an algorithm described herein may account for a movement in a gaze relative to the motion of the head and the object. For example, if the object is moving in an x direction, then effectively the eye gaze may largely not be moving, and eye gaze prediction may be performed.

[0123] In yet another case, if a head motion is largely stationary and eye gazes are continuously moving, then an eye may likely be following a moving object. In such a case, an algorithm described herein may learn an object motion. If the motion of an object is linear, an auto-regression model may be suitable. If the motion of the objectis nonlinear, a neural network may be used to predict the motion of the object. In such a case, head pose may not aid in eye gaze prediction.

[0124] According to a first method described herein, eye gazes may be used at time t-66 ms, t-33 ms, and t ms to predict eye gazes at times t+33 ms, t+66 ms, and t+99ms using an auto regression model. For example, the predictor model 516 may use, as inputs, the head pose data 502, the head pose data 504, the head pose data 506, the eye gaze data 510, the eye gaze data 512, the eye gaze data 514, and the predicted head pose data 508 at time ‘T+t.’ While the predictor model 516 is shown as receiving three eye gaze and head pose data metrics in addition to one predicted head pose metric, the predictor model 516 may have any number of historical eye gaze and head pose metrics, and / or any number of predicted head pose metrics as inputs to determine a predicted eye gaze 518 at time ‘T+ According to a second method, eye gazes and head pose rotation at time t-66 ms, t-33 ms, and t ms may be used to predict eye gazes at times t+33 ms, t+66 ms, and t+99 ms using an auto regression model. According to a baseline method, there may be no prediction of eye gaze. For instance, an eye gaze at time t may be used as a prediction for eye gazes at time t+33 ms, t+66 ms, and t+99 ms. A rotation error may be computed between a ground truth and predicted eye gazes at times t+33 ms, t+66 ms, and t+99 ms for the first method, the second method, and the baseline method. Eye gazes may be available at 30 fps and hence eye gaze samples may be spaced apart by 33 ms.

[0125] FIG. 6 is a diagram 600 illustrating example simulation results 602 and 604 pertaining to a prediction model in accordance with one or more techniques of this disclosure.

[0126] The simulation results may demonstrate that using head pose information along with eye gaze may help in estimating a future eye gaze. Rotation error may be reduced on 99% of errors when head pose information is used. A dataset may simulate an object being mostly stationary. The head may move around in a scene gazing at object without using depth information of the object. Aspects presented herein may pertain to an improved algorithm. To generate the improved algorithm, more data may be generated which may simulate other cases discussed above. The improved algorithm may pertain to a nonlinear predictor (i.e., a nonlinear prediction model) that uses a neural network to model object motion.

[0127] FIG. 7A is a diagram 700 illustrating example aspects pertaining to using XR metadata to enhance an eye gaze prediction in accordance with one or moretechniques of this disclosure. The XR metadata may include depths of objects. Utilizing depths of objects (at which an eye is gazing) may aid in tracking objects in three-dimensional (3D) space. Utilizing the depth of objects may help to leverage a translation component of a head pose. For instance, an eye gaze may be predicted more accurately using rotation and translation components from a head pose. Between frame 702, frame 704, and frame 706, a sensor may track an eye gaze 708 of a user relative to an object 710 displayed in the frames. An eye gaze predictor may calculate a movement 714 of the eye gaze 708, as well as a motion vector 716 relative to the object 710. The delta of the eye gaze 708 between the frame 702, the frame 704, and the frame 706 may be similar to the motion vector 716 in both magnitude and direction. In other words, the user may track the same object between the frame 702, the frame 704, and the frame 706.

[0128] The XR metadata may include motion vectors from an encoder. Aspects presented herein may leverage motion vectors to detect if an eye has started gazing at a different object. If an eye has started to track a different object, then a trajectory of a previous object may lead to inaccuracies (e.g., due to trying to predict a new trajectory from a history of samples of previous trajectories).

[0129] One aspect herein pertains to a method of eye gaze prediction in XR that uses past eye gazes from an eye tracking system as well as head poses from a head tracking system on a device. Obtaining the eye gaze at a future time instance may be based on predicting a head pose at a time instance as a first step. Obtaining the eye gaze at a future time instance may be based on predicting a motion of an object in a space around a user that an eye of the user may be tracking. The object motion may be defined in terms of rotation around a user (e.g., orientation change). Additional cues for determining an object location or a motion of an object may be obtained from metadata / data associated with XR such as depth and motion vectors.

[0130] FIG. 7B is a diagram 750 illustrating example aspects pertaining to using XR metadata to enhance an eye gaze prediction in accordance with one or more techniques of this disclosure. For example, between frame 752, frame 754, and frame 756, a sensor may track an eye gaze 708 of a user relative to an object 710 and an object 712 displayed in the frames. An eye gaze predictor may calculate a movement 714 of the eye gaze 708, as well as a motion vector 716 relative to the object 710. The delta of the eye gaze 708 between the frame 702, the frame 704, and the frame 706 may have adifferent direction when compared to the motion vector 716. In other words, the user may track a different object (e.g., first track the object 710, then track the object 712) between the frame 702, the frame 704, and the frame 706. An eye gaze predictor may then use input samples corresponding to the new object for prediction.

[0131] FIG. 8 is a flowchart 800 illustrating an example workflow to predict a gaze of a user, in accordance with one or more techniques of this disclosure. An eye tracker 802 may capture combined eye gaze directions (e.g., x, y, z vectors) for each eye of a user over time. At 804, the eye gaze predictor may remove samples in an around the user's blink, as such samples may not be as accurate as samples captured at least a threshold time period from the user's blink. For example, a blink may last 200 ms. The eye gaze predictor may remove samples before and after the blink, for example samples within 50 ms before the blink and 50 ms after the blink, or may remove the previous 4 samples before the blink and the latter 4 sample after the blink. In some aspects, the eye gaze predictor may be configured to remove a different number of samples before and after the blink, for example 4 samples before the blink and 2 samples after the blink. The camera may detect a user's blink by tracking the user's eyelid patterns around one or both eyes.

[0132] At 806, the eye gaze predictor may convert the captured eye gazes to a 6D representation. For example, the eye gaze predictor may convert an [x, y, z] gaze direction vector into a 6D rotation. The 6D rotation may illustrate a continuous representation of the rotation. In some aspects, the eye gaze predictor may use a z- vector to obtain an optical axis, where opticalaxis=[0, 0, 1], In some aspects, the eye gaze predictor may use the quaternion headpose [qw, qx, qy, qz] (with respect to the world). In other aspects, the eye gaze predictor may use the eye gaze vector only, for example by setting the head pose to the identity [0, 0, 0, 1] (with respect to the head). The eye gaze predictor may convert the head pose (or identity of the head pose if with respect to the head) to a rotation matrix R. The eye gaze predictor may multiply the rotation matrix with the gaze to obtain a vector product V=inv(R) * gaze. The eye gaze predictor may take the cross-product of the vector product to obtain an axis, where axis = optical axis x V. The eye gaze predictor may calculate the angle, where angle = acos (opticalaxis.V). The eye gaze predictor may convert the axis angle to a rotation matrix having 9 elements. The eye gaze predictor may drop the last column of the 9-element rotation matrix to obtain the 6D representation.

[0133] At 808, the eye gaze predictor may apply a filter to improve the accuracy of the 6D representation input. For example, a causal filter or a low-pass filter. In some aspects, the filter may be an exponential filter, for example a x current_sample + (1-a) x previous_sample.

[0134] At 810, a GRU predictor may receive the 6D filtered input and determine a 6D predicted output based on the historical input data. For example, the GRU predictor described in diagram 900 of FIG. 9. As the GRU predictor receives the filtered 6D input, the GRU predictor may update its hidden states to extrapolate 6D predicted gazes for a number of timestamps in the future. At 812, the eye gaze predictor may convert the 6D predicted output to a gaze direction (e.g., x, y, z vectors) for use in determining the foveated gaze of the user. At 814, the eye gaze predictor may output each predicted gaze to a Tenderer for rendering a frame based on the predicted gaze and the associated timestamp

[0135] FIG. 9 is a diagram 900 illustrating an example aspect of a set of fully connected network (FCN) blocks 902 and a set of gated recurrent unit (GRU) blocks 904 configured to predict a gaze of a user, in accordance with one or more techniques of this disclosure. The set of FCN blocks 902 may be the final layer of an artificial neural network used to predict a set of eye gazes. The set of FCN blocks 902 may be connected to each GRU block of the last layer of the set of GRU blocks 904. The set of GRU blocks may include gating mechanisms for a recurrent neural network. The GRU blocks may accept a 6D continuous representation of the user's gaze to predict a number of 6D predicted outputs of the user's gaze based on an input X1, where X represents information about the gaze at time t. The information about the gaze may include the 6D filtered input from a filter, such as the one at 808 in FIG. 8. The information about the gaze may include at least one of a set of predicted head poses, a field of view (FOV), an estimated depth of an object that the user may be focusing on, and / or a motion vector of a moving obj ect (e.g., a virtual moving obj ect, a physical moving object) that the user may be focusing on from an encoder. The encoder may receive a trajectory of a virtual moving object from an application that generates virtual content for a display. The encoder may determine a trajectory of a physical moving object by analyzing information from a camera that detects objects moving within the FOV of the user. The GRU may be configured to predict a next gaze based on the input X\ In some aspects, a set of GRU blocks 904 may be arranged in a set oflayers to improve the accuracy of the prediction. While four blocks are shown in the set of GRU blocks 904, more or less GRU layers may be used to increase / decrease the accuracy of the prediction. In other words, a four-layer set of GRU blocks may base each time-stamp prediction on at least four previous captured eye gazes. The set of GRU blocks 904 may feed into the set of FCN blocks 902 to provide a set of gaze predictions for future timestamps. The predicted eye gaze of an FCN block may feed into the next set of GRU layers. In other words, if an FCN blocks predicts an eye gaze Yt+1, the predicted eye gaze Yt+1may be fed into an input in the bottom-most GRU layer of the set of GRU blocks 904. The set of FCN blocks 902 may predict eye gazes for 9 future time periods (i.e., Yt+1to Yl 9), or 9 timestamps in the future (TFPs). For example, if each input X1is captured by a camera that captures an eye gaze every 11 ms, then the GRU predictor illustrated in FIG. 9 may predict eye gazes for 99 ms in the future. In some aspects, the GRU predictor may be designed to predict a number of eye gazes greater than or equal to the latency of a round-trip time between transmitting an eye gaze to a server and receiving a rendered frame from the server. For example, if the latency of a round-trip time is 90 ms, then the GRU predictor may be designed to predict a number of eye gazes that is greater or equal to the latency (e.g., 9 predicted eye gazes with a 11 ms period may be used to predict 99 ms worth of eye gazes, which is greater than 90 ms).

[0136] FIG. 10 is a diagram 1000 illustrating an example aspect of a gaze predictor design, in accordance with one or more techniques of this disclosure. An eye tracker input 1002 may include a set of captured eye gazes over time, for example an eye gaze every 11 ms. The eye gaze may be captured as a gaze vector, for example an [x, y, z] direction. The eye tracker input 1002 may include a validity flag. The validity flag may indicate the presence of a blink during a time period. In other words, an eye gaze vector transmitted with a validity flag set may indicate an eye gaze captured without a blink during the time period, and an eye gaze vector transmitted with a validity flag not set may indicate an eye gaze captured with a blink during the time period. The eye tracker input 1002 may include a head pose, for example as a quaternion headpose [qw, qx, qy, qz],

[0137] The eye tracker input 1002 may be received by a blink detector 1006, which may be configured to remove samples in an around the user's blink, as such samples may not be as accurate as samples captured at least a threshold time period from the user'sblink. In other words, the blink detector 1006 may output null values when a captured eye gaze is received during a blink. The blink detector 1006 may transmit a reset flag to the gaze predictor 1018, which may reset the state of any GRUs within a threshold time period of a detected blink. For example, the gaze predictor 1018 may erase GRU state information associated with the previous four eye gaze poses, or may rewind the state of a GRU by four eye gaze poses. In other words, the blink detector 1006 may be configured to erase data associated with a threshold number of eye gaze poses captured before the detected blink. The blink detector 1006 may also be configured to ignore captured eye gaze poses captured shortly after a blink. For example, the blink detector 1006 may output null values for captured eye gaze poses for a threshold of time after a detected blink (e.g., the next 2 eye gaze poses or the next 4 eye gazes).

[0138] The blink detector 1006 may output the filtered gaze directions to a converter 1008 to convert the filtered gaze directions into 6D with respect to the head pose. In other words, the converter 1008 may set the head pose to the identity [0, 0, 0, 1] when calculating a rotation matrix R. The converter 1008 may filter the 6D representation using a low pass filter 1010, which may then be fed to a sequence detector 1016 to determine a sequence of the gaze directions.

[0139] The blink detector 1006 may output the filtered gaze directions to a converter 1012 to convert the filtered gaze directions into 6D with respect to the absolute world about the user. In other words, the converter 1008 may use the quaternion headpose [qw, qx, qy, qz] when calculating a rotation matrix R. The converter 1012 may filter the 6D representation using a low pass filter 1014, which may then be fed to a sequence detector 1016 to determine a sequence of the gaze directions.

[0140] The sequence detector 1016 may detect a state of the eye gaze of a user based on patterns detected by the 6D representation. For example, the sequence detector 1016 may determine that the eye gaze of the user is in a tracking state, for example tracking a moving object. In some aspects, the user may track a virtual moving object. The sequence detector 1016 may determine that the movement vector of the user matches (within a margin of error) a movement vector of a virtual object moving within the FOV of the user. Where the sequence detector 1016 determines that the eye gaze of the user is in a tracking state, the sequence detector 1016 may indicate such to the XOR block 1028, which may then allow the 6D representation with respect to the head of the user from the low pass filter 1010 to be output to the gaze predictor 1018.In another example, the sequence detector 1016 may determine that the eye gaze of the user is in a vestibular ocular reflux (VOR) state, for example tracking a stationary object (e.g., a virtual stationary object or a physical stationary object) or tracking a moving physical object. The sequence detector 1016 may determine that the movement vector of the user matches (within a margin of error) a movement vector of a physical object moving within the FOV of the user. In other words, the user may move their head, while their eye gaze is focused on a stationary virtual or physical object, or the user be tracking a moving physical object. Where the sequence detector 1016 determines that the eye gaze of the user is in a VOR state, the sequence detector 1016 may indicate such to the XOR block 1030, which may then allow the 6D representation with respect to the world of the user from the low pass fdter 1014 to be output to the gaze predictor 1018.

[0141] In some aspects, the sequence detector 1016 may determine that the gaze of the user is in a saccade state. In other words, the sequence detector 1016 may determine that the gaze of the user moves so rapidly that the user isn't focusing on anything. In other words, the measured speed of movement of the eye gaze may be greater than a threshold value. The sequence detector 1016 may output an indication of the saccade to the gaze predictor 1018, which may then reset a state of the GRU. In some aspects, the gaze predictor 1018 may treat a saccade indication the same as a blink, for example by erasing the previous four eye gazes, or by rewinding the state of a set of GRUs by four eye gazes. In other aspects, the gaze predictor 1018 may treat a saccade indication differently than a blink, for example by erasing the previous two eye gazes instead of four. In either case, the gaze predictor 1018 may disable prediction of eye gazes for a period of time in response to a detected saccade. Where the sequence detector 1016 determines that the eye gaze of the user is in a saccade state, the sequence detector 1016 may indicate such to the XOR block 1032, which may then simply pass raw eye gaze data from the low pass filter 1010 to the set of output gazes 1026 — in essence disabling gaze prediction during the saccade. In other words, the sequence detector 1016 may be unable to determine a sequence based on the eye gaze data, which may result in an indeterminate prediction being output as the set of output gazes 1026 for a time period associated with the detected saccade (e.g., during the saccade, for a first threshold number of eye gazes before the detected saccade, and / or for a second threshold number of eye gazes after the detected saccade).

[0142] A gaze predictor 1018 may receive the filtered gaze directions from the blink detector 1006, the determined sequence information from the sequence detector 1016, and / or the filtered 6D representations from the low pass filter 1010 and the low pass filter 1014 to predict a gaze of the user. In some aspects, the gaze predictor 1018 may reset a predictor state of the gaze predictor 1018 in response to the blink detector 1006 detecting a blink. In some aspects, the gaze predictor 1018 may reset a predictor state of the gaze predictor 1018 in response to the sequence detector 1016 detecting a saccade. The gaze predictor 1018 may also receive additional input 1004, for example a set of predicted head poses, a field of view (FOV) of the user, and / or depth / motion vectors from an encoder. In some aspects, the gaze predictor 1018 may ignore / disable data associated with vectors or objects outside of the FOV of the user (e.g., a movement vector of a virtual object that cannot be seen by the user as the virtual object is outside the user's FOV). In some aspects, the gaze predictor 1018 may analyze the combined eye gazes of both of the user's eyes. In some aspects, the gaze predictor 1018 may analyze the eye gaze of each of the user's eyes individually. For example, the gaze predictor 1018 may analyze the eye gaze of each of the user's eyes where the gaze predictor 1018 predicts that the user is following an object having a depth that is less than or equal to a threshold value. In other words, the obj ect (physical or virtual) may be so close to the user's eyes that the user may be cross-eyed, and the gazes of the left and right eyes of the user may be opposite from one another. The gaze predictor 1018 may output a predicted gaze as the set of output gazes 1026, which may include x, y, z gaze vectors at different timestamps (e.g., the next 9 predicted eye gazes).

[0143] FIG. 11 is a connection flow diagram 1100 illustrating an example of a server configured to provide a predicted gaze to a client, in accordance with one or more techniques of this disclosure. The client 1102 may include an HMU. The server may include a server that renders XR content for displaying by the client 1102. The client 1102 and the server 1104 may collectively be a split XR system, where the client 1102 may transmit HMD pose data to the server 1104, and the server 1104 renders XR content to the client 1102, along with the render pose information. The HMD pose data may include head pose data and / or eye gaze data. The server 1104 may compress rendered data before transmitting the rendered data to the client 1102. The client 1102may de-compress and render the rendered frame after warping the frame based on the latest render pose information of the user obtained by the client 1102.

[0144] At 1108, the client 1102 may collect head pose and eye gaze data from a set of sensors monitoring a user, for example sensors of an HMU. The sensors may include, for example, a camera that monitors one or both eyes of a user wearing the HMU, or an IMU that monitors a head pose of a user wearing the HMU. The head pose data may include head pose time series data. The eye gaze data may include eye gaze time series data. The sensors that monitor the head pose and eye gaze information of the user may produce data at different sampling rates. For example, an IMU may generate head pose data at a first sampling rate, a camera may generate eye gaze data at a second sampling rate, and the first sampling rate may be greater than the second sampling rate. In some aspects, the client 1102 may correlate eye gaze data and head pose data by matching timestamps. In other aspects, the client 1102 may transmit raw eye gaze data and head pose data to the server 1104, for the server 1104 to correlate eye gaze data and head pose data by matching timestamps. In other words, a subset of the head pose data may correlate with the set of eye gaze data, a subset of the eye gaze data may correlate with the set of head pose data, or a subset of the eye gaze data may correlate with a subset of the head pose data. The head pose data and / or the eye gaze data may include translation information and orientation information. The eye gaze data may include eye gaze data collected from each eye of the user individually, or a combination of both eye gazes (e.g., an average of the estimated x, y, z values).

[0145] The client 1102 may transmit an indication of the collected data 1110 to a server 1104 for processing. The server 1104 may receive the indication of the collected data 1110 from the client 1102. In some aspects, the client 1102 and the server 1104 may have wireless transceivers that transmit and receive such data, respectively. In some aspects, the collected data 1110 may include a set of predicted head poses of the user. In other aspects, the server 1104 may generate a set of predicted head poses of the user.

[0146] At 1112, the server 1104 may generate a set of predicted eye gazes based on the collected data 1110 received from the client 1102. The server 1104 may generate the set of predicted eye gazes based on historical eye gaze data with respect to the head. The server 1104 may generate the set of predicted eye gazes based on historical eye gaze data and historical head pose data of the user. The server 1104 may generate theset of predicted eye gazes based on XR metadata, such as a depth of a physical / virtual object, a motion vector of the object, and / or a FOV of the user. The server 1104 may generate the set of predicted eye gazes based on other data, such as predicted head pose data. The server 1104 may generate the set of predicted eye gazes based on a predictor model, such as a regression model, a time series forecaster model, or a neural network. In some aspects, the server 1104 may user the workflow of FIG. 8, the FCN blocks and GRU blocks of FIG. 9, and / or the gaze predictor of FIG. 10 to generate the set of predicted eye gazes.

[0147] At 1114, the server 1104 may render a set of frames based on the set of predicted eye gazes. The server 1104 may render a first region of a frame with a higher quality level than a second region of the frame based on a prediction that a user's predicted eye gaze will be in the first region during a time that corresponds with the predicted eye gaze.

[0148] The server 1104 may transmit an indication of the rendered set of frames 1116 to the client 1102. The client 1102 may receive the indication of the rendered set of frames 1116 from the server 1104. In some aspects, the server 1104 may encode the rendered set of frames before transmission, and the client 1102 may decode the encoded rendered set of frames after reception. At 1118, the client 1102 may output the rendered frame, for example to a display of the user. In some aspects, the client 1102 may warp a rendered frame based on a current head pose and a current eye gaze of the user. The current head pose and the current eye gaze of the user may match the predicted eye gaze and / or predicted head pose of the user matching the current timestamp.

[0149] FIG. 12 is a connection flow diagram 1200 illustrating an example of a server configured to provide a predicted gaze to a client, in accordance with one or more techniques of this disclosure. The client 1202 may include an HMU. The server may include a server that renders XR content for displaying by the client 1202. The client 1202 and the server 1204 may collectively be a split XR system, where the client 1202 may transmit HMD pose data to the server 1204, and the server 1204 renders XR content to the client 1202, along with the render pose information. The HMD pose data may include head pose data and / or eye gaze data. The server 1204 may compress rendered data before transmitting the rendered data to the client 1202. The client 1202may de-compress and render the rendered frame after warping the frame based on the latest render pose information of the user obtained by the client 1202.

[0150] At 1208, the client 1202 may collect head pose and eye gaze data from a set of sensors monitoring a user, for example sensors of an HMU. The sensors may include, for example, a camera that monitors one or both eyes of a user wearing the HMU, or an IMU that monitors a head pose of a user wearing the HMU. The head pose data may include head pose time series data. The eye gaze data may include eye gaze time series data. The sensors that monitor the head pose and eye gaze information of the user may produce data at different sampling rates. For example, an IMU may generate head pose data at a first sampling rate, a camera may generate eye gaze data at a second sampling rate, and the first sampling rate may be greater than the second sampling rate. In some aspects, the client 1202 may correlate eye gaze data and head pose data by matching timestamps. In other words, a subset of the head pose data may correlate with the set of eye gaze data, a subset of the eye gaze data may correlate with the set of head pose data, or a subset of the eye gaze data may correlate with a subset of the head pose data. The head pose data and / or the eye gaze data may include translation information and orientation information. The eye gaze data may include eye gaze data collected from each eye of the user individually, or a combination of both eye gazes (e.g., an average of the estimated x, y, z values). In some aspects, the client 1202 may predict a set of predicted head poses of the user.

[0151] At 1212, the client 1202 may generate a set of predicted eye gazes based on the collected data. The client 1202 may generate the set of predicted eye gazes based on historical eye gaze data with respect to the head. The client 1202 may generate the set of predicted eye gazes based on historical eye gaze data and historical head pose data of the user. The client 1202 may generate the set of predicted eye gazes based on XR metadata, such as a depth of a physical / virtual object, a motion vector of the object, and / or a FOV of the user. The server 1104 may generate the set of predicted eye gazes based on other data, such as predicted head pose data. The client 1202 may generate the set of predicted eye gazes based on a predictor model, such as a regression model, a time series forecaster model, or a neural network. In some aspects, the client 1202 may user the workflow of FIG. 8, the FCN blocks and GRU blocks of FIG. 9, and / or the gaze predictor of FIG. 10 to generate the set of predicted eye gazes.

[0152] The client 1202 may transmit an indication of the generated prediction data 1213 to a server 1204 for processing. The server 1204 may receive the indication of the generated prediction data 1213 from the client 1202. In some aspects, the client 1202 and the server 1204 may have wireless transceivers that transmit and receive such data, respectively. In some aspects, the generated prediction data 1213 may include a set of predicted eye gazes of a user, predicted head poses of a user, and / or XR metadata.

[0153] At 1214, the server 1204 may render a set of frames based on the set of predicted eye gazes. The server 1204 may render a first region of a frame with a higher quality level than a second region of the frame based on a prediction that a user's predicted eye gaze will be in the first region during a time that corresponds with the predicted eye gaze.

[0154] The server 1204 may transmit an indication of the rendered set of frames 1216 to the client 1202. The client 1202 may receive the indication of the rendered set of frames 1216 from the server 1204. In some aspects, the server 1204 may encode the rendered set of frames before transmission, and the client 1202 may decode the encoded rendered set of frames after reception. At 1218, the client 1202 may output the rendered frame, for example to a display of the user. In some aspects, the client 1202 may warp a rendered frame based on a current head pose and a current eye gaze of the user. The current head pose and the current eye gaze of the user may match the predicted eye gaze and / or predicted head pose of the user matching the current timestamp.

[0155] FIG. 13 is a flowchart 1300 of an example method of graphics processing in accordance with one or more techniques of this disclosure. The method may be performed by an apparatus, such as an apparatus for graphics processing, a graphics processor (e.g., a GPU), a CPU, the device 104, a client device, a wireless communication device, and the like, as used in connection with the aspects of FIGs. 1-6,7A, 7B, and 8-11. In an example, the method (including the various aspects detailed below) may be performed by the eye gaze predictor 198.

[0156] At 1302, the apparatus (e.g., a client device) may provide, at a first time instance, a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user. In an example, 1302 may be performed by the eye gaze predictor 198. In another example, 1302 may be performed by the client1102 in FIG. 11, which may provide, at a first time instance, an indication of the collected data 1110 to the server 1104. The collected data 1110 may include a first indication of a set of head poses of a head of a user and a second indication of a set of eye poses of an eye of the user collected at 1108. In another example, 1302 may be performed by the client 1202 in FIG. 12, which may provide, at a first time instance, an indication of collected data at 1203 an eye gaze predictor 198. The collected data may include information on a set of head poses of a head of a user and a second indication of a set of eye poses of an eye of the user collected at 1208.

[0157] At 1304, the apparatus (e.g., a client device) obtains, based on the first indication and the second indication, a rendered frame that is based on a predicted eye gaze of the user at a second time instance that occurs after the first time instance. In an example, 1304 may be performed by the eye gaze predictor 198. In another example, 1304 may be performed by the client 1102 in FIG. 11, which may receive an indication of the rendered set of frames 1116 from the server 1104. The client 1102 may receive the indication of the rendered set of frames 1116 based on the first indication and the second indication. At least one rendered frame of the rendered set of frames 1116 may be based on a predicted eye pose of the user at a second time instance that occurs after the first time instance. In another example, 1304 may be performed by the client 1202 in FIG. 12, which may receive the indication of the rendered set of frames 1216 based on the generated prediction data 1213. The generated prediction data 1213 may be based on the first indication and the second indication at 1212. At least one rendered frame of the rendered set of frames 1116 may be based on a predicted eye pose of the user at a second time instance that occurs after the first time instance (e g., at least one of the generated prediction data 1213).

[0158] At 1306, the apparatus (e g., a client device) outputs an indication of the rendered frame. In an example, 1306 may be performed by the eye gaze predictor 198. In another example, 1306 may be performed by the client 1102 in FIG. 11, which may output an indication of the rendered frame. In another example, 1306 may be performed by the client 1202 in FIG. 12, which may output an indication of the rendered frame.

[0159] In one aspect, providing the first indication and the second indication may include transmitting, to a server at the first time instance, the first indication and the secondindication, and where obtaining the rendered frame may include receiving, from the server, the rendered frame.

[0160] In one aspect, outputting the indication of the rendered frame may include (1) storing the rendered frame in at least one of a memory, a buffer, or a cache or (2) transmitting the rendered frame for display.

[0161] In one aspect, the set of head poses may include head pose time series data, and where the set of eye gazes may include eye gaze time series data.

[0162] In one aspect, the apparatus (e.g., a client device) may generate, via an inertial measurement unit (IMU), the set of head poses.

[0163] In one aspect, the apparatus (e.g., a client device) may generate, via a camera, the set of eye gazes.

[0164] In one aspect, generating the set of head poses may include generating the set of head poses at a first sampling rate, where generating the set of eye gazes may include generating the set of eye gazes at a second sampling rate, where the first sampling rate may be greater than the second sampling rate.

[0165] In one aspect, the rendered frame may be encoded, and the apparatus (e.g., a client device) may decode the rendered frame.

[0166] In one aspect, the apparatus (e.g., a client device) may warp the decoded rendered frame based on a latest available head pose of the user, where outputting the indication of the rendered frame may include outputting an indication of the decoded, warped rendered frame.

[0167] In one aspect, the set of head poses may include first translation information and first orientation information of the head of the user, and where the set of eye gazes may include second translation information and second orientation information of the eye of the user.

[0168] In one aspect, the rendered frame may include a first region and a second region, where the first region may be associated with a first quality level and the second region may be associated with a second quality level, where the first quality level may be greater than the second quality level, and where the first region may correspond to the predicted eye gaze of the user at the second time instance.

[0169] In one aspect, the set of head poses may correspond to first time instances occurring prior to or concurrently with the first time instance, and where the set of eye gazesmay correspond to second time instances occurring prior to or concurrently with the first time instance.

[0170] In one aspect, the set of eye gazes of the user may correspond to eye gazes of the eye of the user on a display, and where the predicted eye gaze of the user may correspond to a predicted eye gaze of the user on the display.

[0171] In one aspect, the rendered frame may be further based on extended reality (XR) metadata including at least one of depth information associated with an object or a motion vector associated with the object, and where the object may be a virtual object or a real -world object.

[0172] FIG. 14 is a flowchart 1400 of an example method of graphics processing in accordance with one or more techniques of this disclosure. The method may be performed by an apparatus, such as an apparatus for graphics processing, a graphics processor (e g., a GPU), a CPU, the device 104, a server, a wireless communication device, and the like, as used in connection with the aspects of FIGs. 1-6, 7A, 7B, and 8-11. In an example, the method (including the various aspects detailed below) may be performed by the eye gaze predictor 199.

[0173] At 1402, the apparatus (e.g., a server) obtains a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user, where the first indication and the second indication correspond to a first time instance. In an example, 1402 may be performed by the eye gaze predictor 199. In another example, 1402 may be performed by the server 1104 in FIG. 11, which may receive the collected data 1110 from the client 1102. The collected data 1110 may include a first indication of a set of head poses of a head of a user and a second indication of a set of eye poses of an eye of the user. The first indication and the second indication may correspond to a first time instance. In another example, 1402 may be performed by the client 1202 in FIG. 12, which may obtain a first indication of a set of head poses of a head of a user and a second indication of a set of eye poses of an eye of the user collected at 1208. The first indication and the second indication may correspond to a first time instance.

[0174] At 1404, the apparatus (e.g., a server) generates, based on the first indication and the second indication, a predicted eye gaze of the user, where the predicted eye gaze of the user corresponds to a second time instance that occurs after the first time instance. In an example, 1404 may be performed by the eye gaze predictor 199. In anotherexample, 1404 may be performed by the server 1104 in FIG. 11, which may, at 1112, generate, based on the first indication and the second indication, a predicted eye pose of the user. The predicted eye pose of the user may correspond to a second time instance that occurs after the first time instance. In another example, 1404 may be performed by the client 1202 in FIG. 12, which may, at 1212, generate, based on the first indication and the second indication, a predicted eye pose of the user. The predicted eye pose of the user may correspond to a second time instance that occurs after the first time instance.

[0175] At 1406, the apparatus (e.g., a server) renders a frame based on the predicted eye gaze of the user. In an example, 1406 may be performed by the eye gaze predictor 199. In another example, 1406 may be performed by the server 1104 in FIG. 11, which may, at 1114, render a frame based on the predicted eye pose of the user. In another example, 1406 may be performed by the client 1202 in FIG. 12, which may render a frame based on the predicted eye pose of the user.

[0176] At 1408, the apparatus (e.g., a server) outputs the rendered frame. In an example, 1408 may be performed by the eye gaze predictor 199. In another example, 1408 may be performed by the server 1104 in FIG. 11, which may transmit the rendered set of frames 1216 to the client 1202. In another example, 1408 may be performed by the client 1202 in FIG. 12, which may, at 1218, output the rendered frame.

[0177] FIG. 15 is a flowchart 1500 of an example method of graphics processing in accordance with one or more techniques of this disclosure. The method may be performed by an apparatus, such as an apparatus for graphics processing, a graphics processor (e.g., a GPU), a CPU, the device 104, the client 1102, the server 1104, a server, a wireless communication device, and the like, as used in connection with the aspects of FIGs. 1-6, 7A, 7B, and 8-11. In an example, the method (including the various aspects detailed below) may be performed by the eye gaze predictor 199.

[0178] At 1502, the apparatus may obtain a first indication of a first set of eye gazes associated with a user device. For example, 1502 may be performed by the server 1104 in FIG. 11, which may receive the collected data 1110 from the client 1102. The collected data 1110 may include a first indication of a first set of eye gazes associated with a user device. In another example, 1502 may be performed by the client 1202 in FIG. 12, which may receive, at 1208, a first indication of a first set of eye gazes associated with a user device. In another example, 1502 may be performed by the eyegaze predictor 198. In another example, 1502 may be performed by the eye gaze predictor 199.

[0179] At 1504, the apparatus may obtain a second indication of a second set of head poses associated with the user device. For example, 1504 may be performed by the server 1104 in FIG. 11, which may receive the collected data 1110 from the client 1102. The collected data 1110 may include a second indication of a second set of head poses associated with the user device. In another example, 1504 may be performed by the client 1202 in FIG. 12, which may obtain, at 1208, a second indication of a second set of head poses associated with the user device. In another example, 1504 may be performed by the eye gaze predictor 198. In another example, 1504 may be performed by the eye gaze predictor 199.

[0180] At 1506, the apparatus may determine a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes. For example, 1506 may be performed by the server 1104 in FIG. 11, which may, at 1112, determine a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes. In another example, 1506 may be performed by the client 1202 in FIG. 12, which may, at 1212, determine a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes. In another example, 1506 may be performed by the eye gaze predictor 199. In another example, 1506 may be performed by the eye gaze predictor 198. In another example, 1506 may be performed by the eye gaze predictor 199.

[0181] At 1508, the apparatus may output a third indication of the determined third set of predicted eye gazes. For example, 1508 may be performed by the server 1104 in FIG. 11, which may output a third indication of the determined third set of predicted eye gazes to a Tenderer at 1114 to render a set of frames based on the predicted eye gazes. In another example, 1508 may be performed by the client 1202 in FIG. 12, which may transmit an indication of the generated prediction data 1213 to the server 1204. The generated prediction data 1213 may include a third indication of the determined third set of predicted eye gazes. In another example, 1508 may be performed by the eyegaze predictor 199. In another example, 1508 may be performed by the eye gaze predictor 199.

[0182] In one aspect, obtaining the first indication and the second indication may include receiving, from a client device, the first indication and the second indication, and where outputting the rendered frame may include transmitting, to the client device, the rendered frame.

[0183] In one aspect, the set of head poses may include head pose time series data, and where the set of eye gazes may include eye gaze time series data.

[0184] In one aspect, the set of head poses may be associated with a first sampling rate and the set of eye gazes may be associated with a second sampling rate, and where the first sampling rate may be greater than the second sampling rate.

[0185] In one aspect, the apparatus may encode the rendered frame, where outputting the rendered frame may include transmitting the encoded rendered frame.

[0186] In one aspect, the set of head poses may include first translation information and first orientation information of the head of the user, and where the set of eye gazes may include second translation information and second orientation information of the eye of the user.

[0187] In one aspect, the rendered frame may include a first region and a second region, where the first region may be associated with a first quality level and the second region may be associated with a second quality level, where the first quality level may be greater than the second quality level, and where the first region may correspond to the predicted eye gaze of the user at the second time instance.

[0188] In one aspect, the set of head poses may correspond to first time instances occurring prior to or concurrently with the first time instance, and where the set of eye gazes may correspond to second time instances occurring prior to or concurrently with the first time instance.

[0189] In one aspect, the set of eye gazes of the user may correspond to eye gazes of the eye of the user on a display, and where the predicted eye gaze of the user may correspond to a predicted eye gaze of the user on the display.

[0190] In one aspect, the apparatus may obtain extended reality (XR) metadata, where generating the predicted eye gaze of the user may include generating the predicted eye gaze of the user further based on the XR metadata.

[0191] In one aspect, the XR metadata may include at least one of depth information associated with an object or a motion vector associated with the object.

[0192] In one aspect, the object may be a virtual object or a real -world object.

[0193] In one aspect, generating the predicted eye gaze of the user may include generating the predicted eye gaze of the user via a predictor model.

[0194] In one aspect, the predictor model may include at least one of a regression model, a time series forecaster model, or a neural network.

[0195] In configurations, a method or an apparatus for graphics processing is provided. The apparatus may be a GPU, a CPU, or some other processor that may perform graphics processing. In aspects, the apparatus may be the processing unit 120 within the device 104, or may be some other hardware within the device 104 or another device. The apparatus may include means for obtaining an indication of an eye gaze associated with a user device. The apparatus may include means for obtaining an indication of a head pose associated with the user device. The apparatus may include means for determining a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose. The apparatus may include means for outputting an indication of the determined set of predicted eye gazes. The apparatus may include means for correlating the head pose with the eye gaze. The apparatus may include means for correlating the head pose with the eye gaze by correlating the head pose with the eye gaze based on a first time indicator associated with the eye gaze and a second time indicator associated with the head pose. The indication of the eye gaze may include eye gaze time series data. The indication of the head pose may include head pose time series data. The apparatus may include means for obtaining an indication of a second eye gaze associated with the user device. The apparatus may include means for obtaining an indication of a set of blinks associated with the user device. The apparatus may include means for determining that the second eye gaze is associated with at least one of the set of blinks. The apparatus may include means for determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose by refraining from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of blinks. The apparatus may include means for obtaining an indication of a second eye gaze associated with the user device. The apparatus may includemeans for obtaining an indication of a set of saccades associated with the user device. The apparatus may include means for determining that the second eye gaze is associated with at least one of the set of saccades. The apparatus may include means for determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose by refraining from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of saccades. The apparatus may include means for converting the eye gaze into a 6D representation of the eye gaze. The apparatus may include means for determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose by determining the set of predicted eye gazes based on the converted 6D representation of the eye gaze. The apparatus may include means for determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose by determining the set of predicted eye gazes based on a first set of FCN blocks and a set of GRU blocks. The apparatus may include means for determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose by determining the set of predicted eye gazes based on a predictor model that uses the eye gaze and the head pose as inputs. The predictor model may include at least one of a regression model, a time series forecaster model, or a neural network. The apparatus may include means for obtaining an indication of a second eye gaze associated with the user device. The apparatus may include means for determining that the second eye gaze is associated with a user tracking a virtual object. The apparatus may include means for determining a second set of predicted eye gazes based on the obtained indication of the second eye gaze with respect to a head of the user in response to a determination that the eye gaze is associated with the user tracking the virtual object. The apparatus may include means for outputting an indication of the determined second set of predicted eye gazes. The apparatus may include means for determining that the eye gaze is not associated with the user tracking the virtual object. The apparatus may include means for determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose by determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises in response to adetermination that the eye gaze is not associated with the user tracking the virtual object. The apparatus may include means for obtaining the indication of the eye gaze associate with the user device by generating, via an eye gaze camera associated with the user device, a set of eye gazes comprising the eye gaze at a first sampling rate. The apparatus may include means for obtaining the head pose by generating, via an IMU associated with the user device, a set of head poses comprising the head pose at a second sampling rate. The first sampling rate may be greater than the second sampling rate. The indication of the eye gaze may include first translation information and first orientation information of at least one eye of a user associated with the user device. The indication of the head pose may include second translation information and second orientation information of a head of the user associated with the user device. The apparatus may include means for outputting the indication of the determined set of predicted eye gazes by transmitting the indication of the determined set of predicted eye gazes. The apparatus may include means for receiving an indication of a set of rendered frames after a transmission of the indication of the determined set of predicted eye gazes. The apparatus may include means for outputting an indication of the set of rendered frames. The apparatus may include means for outputting the indication of the determined set of predicted eye gazes by (a) transmitting the indication of the determined set of predicted eye gazes, (b) receiving an indication of a set of rendered frames after a transmission of the indication of the determined set of predicted eye gazes and (c) outputting an indication of the set of rendered frames. The apparatus may include means for outputting the indication of the set of rendered frames by storing the set of rendered frames in at least one of a memory, a buffer, or a cache. The apparatus may include means for decoding the set of rendered frames. The apparatus may include means for warping the decoded set of rendered frames based on a latest available head pose of a user associated with the user device. The apparatus may include means for outputting the indication of the set of rendered frames by outputting an indication of a set of warped, decoded rendered frames. Each of the set of rendered frames may include a first region and a second region. The first region may be associated with a first quality level and the second region is associated with a second quality level. The first quality level may be greater than the second quality level. The first region may correspond to at least one of the set of predicted eye gazes. The apparatus may include means for obtainingthe indication of the eye gaze associated with the user device by receiving the indication of the eye gaze from the user device. The apparatus may include means for obtaining the indication of the head pose associated with the user device by receiving the indication of the head pose from the user device. The apparatus may include means for outputting the indication of the determined set of predicted eye gazes by (a) outputting the indication to a Tenderer. The apparatus may include means for rendering, via the Tenderer, a set of frames based on the determined set of predicted eye gazes. The apparatus may include means for transmitting the rendered set of frames to the user device. The apparatus may include means for rendering the set of frames based on the determined set of predicted eye gazes by rendering a first region and a second region of each of the set of frames. The first region may be associated with a first quality level and the second region may be associated with a second quality level. The first quality level may be greater than the second quality level. The first region may correspond with at least one of the determined set of predicted eye gazes. The apparatus may include means for transmitting the rendered set of frames to the user device by (a) encoding the rendered set of frames, and (b) transmitting the encoded rendered set of frames. In some aspects, the means may include the eye gaze predictor 198. In some aspects, the means may include the eye gaze predictor 199.

[0196] The apparatus may include means for providing, at a first time instance, a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user. The apparatus may further include means for obtaining, based on the first indication and the second indication, a rendered frame that is based on a predicted eye gaze of the user at a second time instance that occurs after the first time instance. The apparatus may further include means for outputting an indication of the rendered frame. The apparatus may further include means for generating, via an inertial measurement unit (IMU), the set of head poses. The apparatus may further include means for generating, via a camera, the set of eye gazes. The apparatus may further include means for decoding the rendered frame. The apparatus may further include means for warping the decoded rendered frame based on a latest available head pose of the user, where outputting the indication of the rendered frame includes outputting an indication of the decoded, warped rendered frame.

[0197] In configurations, a method or an apparatus for graphics processing is provided. The apparatus may be a GPU, a CPU, or some other processor that may perform graphics processing. In aspects, the apparatus may be the processing unit 120 within the device 104, or may be some other hardware within the device 104 or another device. The apparatus may include means for obtaining a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user, where the first indication and the second indication correspond to a first time instance. The apparatus may further include means for generating, based on the first indication and the second indication, a predicted eye gaze of the user, where the predicted eye gaze of the user corresponds to a second time instance that occurs after the first time instance. The apparatus may further include means for rendering a frame based on the predicted eye gaze of the user. The apparatus may further include means for outputting the rendered frame. The apparatus may further include means for encoding the rendered frame, where outputting the rendered frame includes transmitting the encoded rendered frame. The apparatus may further include means for obtaining extended reality (XR) metadata, where generating the predicted eye gaze of the user includes generating the predicted eye gaze of the user further based on the XR metadata.

[0198] In configurations, a method or an apparatus for graphics processing is provided. The apparatus may be a GPU, a CPU, or some other processor that may perform graphics processing. In aspects, the apparatus may be the processing unit 120 within the device 104, or may be some other hardware within the device 104 or another device. The apparatus may include means for obtaining a first indication of a first set of eye gazes associated with a user device. The apparatus may include means for obtaining a second indication of a second set of head poses associated with the user device. The apparatus may include means for determining a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes. The apparatus may include means for outputting a third indication of the determined third set of predicted eye gazes. The apparatus may include means for correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes. The apparatus may include means for correlating the second subset of the obtained second set of head poses with the first subset of theobtained first set of eye gazes by correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes based on a first set of time indicators associated with the first set of eye gazes and a second set of time indicators associated with the second set of head poses. The first set of eye gazes may include eye gaze time series data. The second set of head poses may include head pose time series data. The apparatus may include means for obtaining a fourth indication of a fourth set of blinks associated with the user device. The apparatus may include means for selecting a third subset of the obtained first set of eye gazes based on the fourth set of blinks associated with the user device. The apparatus may include means for determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes by determining the third set of predicted eye gazes based on the selected third subset of the obtained first set of eye gazes and a fourth subset of the obtained second set of head poses correlated with the selected third subset of the obtained first set of eye gazes. The apparatus may include means for determining that a third subset of the obtained first set of eye gazes is associated with a saccade. The apparatus may include means for selecting a fourth subset of the obtained first set of eye gazes based on the determined third subset of the obtained first set of eye gazes associated with the saccade. The apparatus may include means for determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes by determining the third set of predicted eye gazes based on the selected fourth subset of the obtained first set of eye gazes and a fifth subset of the obtained second set of head poses correlated with the selected fourth subset of the obtained first set of eye gazes. The apparatus may include means for converting a third subset of the first set of eye gazes into a fourth set of six-dimensional (6D) representations of the first set of eye gazes. The apparatus may include means for determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes by determining the third set of predicted eye gazes based on the converted fourth set of 6D representations of the first set of eye gazes. The apparatus may include means for determining the third set of predicted eye gazes based on theobtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes by determining the third set of predicted eye gazes based on a first set of fully-connected network (FCN) blocks and a set of gated recurrent unit (GRU) blocks. The apparatus may include means for determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes by determining the third set of predicted eye gazes based on a predictor model that uses the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes as inputs. The predictor model may include at least one of a regression model, a time series forecaster model, or a neural network. The apparatus may include means for determining that a third subset of the first set of eye gazes is associated with a user tracking a virtual object. The apparatus may include means for determining a fourth set of predicted eye gazes based on the determined third subset of the first set of eye gazes with respect to a head of the user. The apparatus may include means for outputting a fourth indication of the determined fourth set of predicted eye gazes. The apparatus may include means for determining that a fourth subset of the first set of eye gazes are not associated with the user tracking the virtual object. The apparatus may include means for determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes by determining the third set of predicted eye gazes based on the determined fourth subset of the first set of eye gazes and a fifth subset of the obtained second set of head poses correlated with the determined fourth subset of the first set of eye gazes. The apparatus may include means for obtaining the first set of eye gazes by generating, via an eye gaze camera associated with the user device via, the first set of eye gazes at a first sampling rate. The apparatus may include means for obtaining the second set of head poses by generating, via an inertial measurement unit (IMU) associated with the user device, the second set of head poses at a second sampling rate. The first sampling rate may be greater than the second sampling rate. The first set of eye gazes may include a third set of eye gaze time series data. The second set of head poses may include a fourth set of head pose time series data. The first set of eye gazes may include firsttranslation information and first orientation information of at least one eye of a user associated with the user device. The second set of head poses may include second translation information and second orientation information of a head of the user associated with the user device. The apparatus may include means for outputting the third indication of the determined third set of predicted eye gazes by (a) transmitting the third indication of the determined third set of predicted eye gazes, (b) receiving a fourth indication of a fourth set of rendered frames based on a transmission of the third indication of the determined third set of predicted eye gazes, and (c) outputting a fifth indication of the fourth set of rendered frames. The apparatus may include means for outputting the fifth indication of the fourth set of rendered frames by storing the fourth set of rendered frames in at least one of a memory, a buffer, or a cache. The apparatus may include means for decoding the fourth set of rendered frames; and warping the decoded fourth set of rendered frames based on a latest available head pose of a user associated with the user device. The apparatus may include means for outputting the fifth indication of the fourth set of rendered frames by outputting a sixth indication of a fifth set of warped, decoded rendered frames. Each of the fourth set of rendered frames may include a first region and a second region. The first region may be associated with a first quality level and the second region is associated with a second quality level. The first quality level may be greater than the second quality level. The first region may correspond to at least one of the third set of predicted eye gazes. The apparatus may include means for obtaining the first indication of the first set of eye gazes associated with the user device by receiving the first indication of the first set of eye gazes from the user device. The apparatus may include means for obtaining the second indication of the second set of head poses associated with the user device by receiving the second indication of the second set of head poses from the user device. The apparatus may include means for outputting the third indication of the determined third set of predicted eye gazes by outputting the third indication to a Tenderer. The apparatus may include means for rendering, via the Tenderer, a fourth set of frames based on the determined third set of predicted eye gazes; and transmitting the rendered fourth set of frames to the user device. The apparatus may include means for rendering the fourth set of frames based on the determined third set of predicted eye gazes by rendering a first region and a second region of each of the fourth set of frames. The first region may be associated with a first quality level andthe second region is associated with a second quality level. The first quality level may be greater than the second quality level. The first region may correspond with at least one of the determined third set of predicted eye gazes. The apparatus may include means for transmitting the rendered fourth set of frames to the user device by (a) encoding the rendered fourth set of frames, and (b) transmitting the encoded rendered fourth set of frames.

[0199] It is understood that the specific order or hierarchy of blocks / steps in the processes, flowcharts, and / or call flow diagrams disclosed herein is an illustration of example approaches. Based upon design preferences, it is understood that the specific order or hierarchy of the blocks / steps in the processes, flowcharts, and / or call flow diagrams may be rearranged. Further, some blocks / steps may be combined and / or omitted. Other blocks / steps may also be added. The accompanying method claims present elements of the various blocks / steps in a sample order, and are not meant to be limited to the specific order or hierarchy presented.

[0200] The previous description is provided to enable any person skilled in the art to practice the various aspects described herein. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects. Thus, the claims are not intended to be limited to the aspects shown herein, but is to be accorded the full scope consistent with the language of the claims, where reference to an element in the singular is not intended to mean “one and only one” unless specifically so stated, but rather “one or more.” The word “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any aspect described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects.

[0201] Unless specifically stated otherwise, the term “some” refers to one or more and the term “or” may be interpreted as “and / or” where context does not dictate otherwise. Combinations such as “at least one of A, B, or C,” “one or more of A, B, or C,” “at least one of A, B, and C,” “one or more of A, B, and C,” and “A, B, C, or any combination thereof’ include any combination of A, B, and / or C, and may include multiples of A, multiples of B, or multiples of C. Specifically, combinations such as “at least one of A, B, or C,” “one or more of A, B, or C,” “at least one of A, B, and C,” “one or more of A, B, and C,” and “A, B, C, or any combination thereof’ may be A only, B only, C only, A and B, A and C, B and C, or A and B and C, where anysuch combinations may contain one or more member or members of A, B, or C. All structural and functional equivalents to the elements of the various aspects described throughout this disclosure that are known or later come to be known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be encompassed by the claims. Moreover, nothing disclosed herein is intended to be dedicated to the public regardless of whether such disclosure is explicitly recited in the claims. The words “module,” “mechanism,” “element,” “device,” and the like may not be a substitute for the word “means.” As such, no claim element is to be construed as a means plus function unless the element is expressly recited using the phrase “means for.” Unless stated otherwise, the phrase “a processor” may refer to “any of one or more processors” (e.g., one processor of one or more processors, a number (greater than one) of processors in the one or more processors, or all of the one or more processors) and the phrase “a memory” may refer to “any of one or more memories” (e.g., one memory of one or more memories, a number (greater than one) of memories in the one or more memories, or all of the one or more memories).

[0202] In one or more examples, the functions described herein may be implemented in hardware, software, firmware, or any combination thereof. For example, although the term “processing unit” has been used throughout this disclosure, such processing units may be implemented in hardware, software, firmware, or any combination thereof. If any function, processing unit, technique described herein, or other module is implemented in software, the function, processing unit, technique described herein, or other module may be stored on or transmitted over as one or more instructions or code on a computer-readable medium.

[0203] Computer-readable media may include computer data storage media or communication media including any medium that facilitates transfer of a computer program from one place to another. In this manner, computer-readable media generally may correspond to: (1) tangible computer-readable storage media, which is non-transitory; or (2) a communication medium such as a signal or carrier wave. Data storage media may be any available media that can be accessed by one or more computers or one or more processors to retrieve instructions, code, and / or data structures for implementation of the techniques described in this disclosure. By way of example, and not limitation, such computer-readable media may include RAM,ROM, EEPROM, compact disc-read only memory (CD-ROM), or other optical disk storage, magnetic disk storage, or other magnetic storage devices. Disk and disc, as used herein, includes compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk, and Blu-ray disc, where disks usually reproduce data magnetically, while discs usually reproduce data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media. A computer program product may include a computer-readable medium.

[0204] The techniques of this disclosure may be implemented in a wide variety of devices or apparatuses, including a wireless handset, an integrated circuit (IC) or a set of ICs, e.g., a chip set. Various components, modules or units are described in this disclosure to emphasize functional aspects of devices configured to perform the disclosed techniques, but do not necessarily need realization by different hardware units. Rather, as described above, various units may be combined in any hardware unit or provided by a collection of inter-operative hardware units, including one or more processors as described above, in conjunction with suitable software and / or firmware. Accordingly, the term “processor,” as used herein may refer to any of the foregoing structure or any other structure suitable for implementation of the techniques described herein. Also, the techniques may be fully implemented in one or more circuits or logic elements.

[0205] The following aspects are illustrative only and may be combined with other aspects or teachings described herein, without limitation.

[0206] Aspect 1 is a method of graphics processing, comprising: providing, at a first time instance, a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user; obtaining, based on the first indication and the second indication, a rendered frame that is based on a predicted eye gaze of the user at a second time instance that occurs after the first time instance; and; and outputting an indication of the rendered frame.

[0207] Aspect 2 is the method of aspect 1, wherein providing the first indication and the second indication comprises transmitting, to a server at the first time instance, the first indication and the second indication, and wherein obtaining the rendered frame comprises receiving from the server, the rendered frame.

[0208] Aspect 3 is the method of aspect 2, wherein the graphics processing is performed on a wireless communication device comprising at least one of a transceiver or anantenna, wherein receiving the rendered frame comprises receiving the rendered frame via at least one of the transceiver or the antenna.

[0209] Aspect 4 is the method of any of aspects 1 to 3, wherein outputting the indication of the rendered frame comprises: (1) storing the rendered frame in at least one of the memory, a buffer, or a cache or (2) transmitting the rendered frame for display.

[0210] Aspect 5 is the method of any of aspects 1 to 4, wherein the set of head poses comprises head pose time series data, and wherein the set of eye gazes comprises eye gaze time series data.

[0211] Aspect 6 is the method of any of aspects 1 to 5, further comprising: generating, via an inertial measurement unit (IMU), the set of head poses; and generating, via a camera, the set of eye gazes.

[0212] Aspect 7 is the method of aspect 6, wherein generating the set of head poses comprises generating the set of head poses at a first sampling rate, wherein generating the set of eye gazes comprises generating the set of eye gazes at a second sampling rate, wherein the first sampling rate is less than the second sampling rate.

[0213] Aspect 8 is the method of any of aspects 1 to 7, wherein the rendered frame is encoded, further comprising: decoding the rendered frame; and warping the decoded rendered frame based on a latest available head pose of the user, wherein outputting the indication of the rendered frame comprises outputting an indication of the decoded, warped rendered frame.

[0214] Aspect 9 is the method of any of aspects 1 to 8, wherein the set of head poses comprises first translation information and first orientation information of the head of the user, and wherein the set of eye gazes comprises second translation information and second orientation information of the eye of the user.

[0215] Aspect 10 is the method of any of aspects 1 to 9, wherein the rendered frame comprises a first region and a second region, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds to the predicted eye gaze of the user at the second time instance.

[0216] Aspect 11 is the method of any of aspects 1 to 10, wherein the set of head poses corresponds to first time instances occurring prior to or concurrently with the firsttime instance, and wherein the set of eye gazes corresponds to second time instances occurring prior to or concurrently with the first time instance.

[0217] Aspect 12 is the method of any of aspects 1 to 11, wherein the set of eye gazes of the user corresponds to eye gazes of the eye of the user on a display, and wherein the predicted eye gaze of the user corresponds to a predicted eye gaze of the user on the display.

[0218] Aspect 13 is the method of any of aspects 1 to 12, wherein the rendered frame is further based on extended reality (XR) metadata comprising at least one of depth information associated with an object or a motion vector associated with the object, and wherein the object is a virtual object or a real-world object.

[0219] Aspect 14 is a method of graphics processing, comprising: obtaining a first indication of a set of head poses of a head of a user and a second indication of a set of eye gazes of an eye of the user, wherein the first indication and the second indication correspond to a first time instance; generating, based on the first indication and the second indication, a predicted eye gaze of the user, wherein the predicted eye gaze of the user corresponds to a second time instance that occurs after the first time instance; rendering a frame based on the predicted eye gaze of the user; and outputting the rendered frame.

[0220] Aspect 15 is the method of aspect 14, wherein obtaining the first indication and the second indication comprises: receiving, from a client device, the first indication and the second indication, and wherein outputting the rendered frame comprises transmitting, to the client device, the rendered frame.

[0221] Aspect 16 is the method of aspect 15, wherein the method is performed by a wireless communication device comprising at least one of a transceiver or an antenna, wherein transmitting the rendered frame comprises transmitting the rendered frame via at least one of the transceiver or the antenna.

[0222] Aspect 17 is the method of any of aspects 14 to 16, wherein the set of head poses comprises head pose time series data, and wherein the set of eye gazes comprises eye gaze time series data.

[0223] Aspect 18 is the method of any of aspects 14 to 17, wherein the set of head poses is associated with a first sampling rate and the set of eye gazes is associated with a second sampling rate, and wherein the first sampling rate is less than the second sampling rate.

[0224] Aspect 19 is the method of any of aspects 14 to 18, further comprising: encoding the rendered frame, wherein outputting the rendered frame comprises transmitting the encoded rendered frame.

[0225] Aspect 20 is the method of any of aspects 14 to 19, wherein the set of head poses comprises first translation information and first orientation information of the head of the user, and wherein the set of eye gazes comprises second translation information and second orientation information of the eye of the user.

[0226] Aspect 21 is the method of any of aspects 14 to 20, wherein the rendered frame comprises a first region and a second region, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds to the predicted eye gaze of the user at the second time instance

[0227] Aspect 22 is the method of any of aspects 14 to 21, wherein the set of head poses corresponds to first time instances occurring prior to or concurrently with the first time instance, and wherein the set of eye gazes corresponds to second time instances occurring prior to or concurrently with the first time instance.

[0228] Aspect 23 is the method of any of aspects 14 to 22, wherein the set of eye gazes of the user corresponds to eye gazes of the eye of the user on a display, and wherein the predicted eye gaze of the user corresponds to a predicted eye gaze of the user on the display.

[0229] Aspect 24 is the method of any of aspects 14 to 23, further comprising: obtaining extended reality (XR) metadata, wherein generating the predicted eye gaze of the user comprises generating the predicted eye gaze of the user further based on the XR metadata.

[0230] Aspect 25 is the method of aspect 24, wherein the XR metadata comprises at least one of depth information associated with an object or a motion vector associated with the object.

[0231] Aspect 26 is the method of aspect 25, wherein the object is a virtual object or a real- world object.

[0232] Aspect 27 is the method of any of aspects 14 to 26, wherein generating the predicted eye gaze of the user comprises generating the predicted eye gaze of the user via a predictor model.

[0233] Aspect 28 is the method of aspect 27, wherein the predictor model comprises at least one of a regression model, a time series forecaster model, or a neural network.

[0234] Aspect 29 is a method of graphics processing, comprising: obtaining a first indication of a first set of eye gazes associated with a user device; obtaining a second indication of a second set of head poses associated with the user device; determining a third set of predicted eye gazes based on the obtained first set of eye gazes and a second subset of the obtained second set of head poses correlated with a first subset of the obtained first set of eye gazes; and outputting a third indication of the determined third set of predicted eye gazes.

[0235] Aspect 30 is the method of aspect 29, further comprising: correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes.

[0236] Aspect 31 is the method of aspect 30, wherein correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes comprises: correlating the second subset of the obtained second set of head poses with the first subset of the obtained first set of eye gazes based on a first set of time indicators associated with the first set of eye gazes and a second set of time indicators associated with the second set of head poses.

[0237] Aspect 32 is the method of any of aspects 29 to 31, wherein the first set of eye gazes comprises eye gaze time series data, wherein the second set of head poses comprises head pose time series data.

[0238] Aspect 33 is the method of any of aspects 29 to 32, further comprising: obtaining a fourth indication of a fourth set of blinks associated with the user device; and selecting a third subset of the obtained first set of eye gazes based on the fourth set of blinks associated with the user device, wherein determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes comprises: determining the third set of predicted eye gazes based on the selected third subset of the obtained first set of eye gazes and a fourth subset of the obtained second set of head poses correlated with the selected third subset of the obtained first set of eye gazes.

[0239] Aspect 34 is the method of any of aspects 29 to 33, further comprising: determining that a third subset of the obtained first set of eye gazes is associated with a saccade;and selecting a fourth subset of the obtained first set of eye gazes based on the determined third subset of the obtained first set of eye gazes associated with the saccade, wherein determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes comprises: determining the third set of predicted eye gazes based on the selected fourth subset of the obtained first set of eye gazes and a fifth subset of the obtained second set of head poses correlated with the selected fourth subset of the obtained first set of eye gazes.

[0240] Aspect 35 is the method of any of aspects 29 to 34, further comprising: converting a third subset of the first set of eye gazes into a fourth set of six-dimensional (6D) representations of the first set of eye gazes, wherein determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes comprises: determining the third set of predicted eye gazes based on the converted fourth set of 6D representations of the first set of eye gazes.

[0241] Aspect 36 is the method of any of aspects 29 to 35, wherein determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes comprises: determining the third set of predicted eye gazes based on a first set of fully-connected network (FCN) blocks and a set of gated recurrent unit (GRU) blocks.

[0242] Aspect 37 is the method of any of aspects 29 to 36, wherein determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes comprises: determining the third set of predicted eye gazes based on a predictor model that uses the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes as inputs.

[0243] Aspect 38 is the method of aspect 37, wherein the predictor model comprises at least one of a regression model, a time series forecaster model, or a neural network.

[0244] Aspect 39 is the method of any of aspects 29 to 38, further comprising: determining that a third subset of the first set of eye gazes is associated with a user tracking a virtual object; determining a fourth set of predicted eye gazes based on the determinedthird subset of the first set of eye gazes with respect to a head of the user; and outputting a fourth indication of the determined fourth set of predicted eye gazes.

[0245] Aspect 40 is the method of aspect 39, further comprising: determining that a fourth subset of the first set of eye gazes are not associated with the user tracking the virtual object, wherein determining the third set of predicted eye gazes based on the obtained first set of eye gazes and the second subset of the obtained second set of head poses correlated with the first subset of the obtained first set of eye gazes comprises: determining the third set of predicted eye gazes based on the determined fourth subset of the first set of eye gazes and a fifth subset of the obtained second set of head poses correlated with the determined fourth subset of the first set of eye gazes.

[0246] Aspect 41 is the method of any of aspects 29 to 40, wherein obtaining the first set of eye gazes comprises: generating, via an eye gaze camera associated with the user device via, the first set of eye gazes at a first sampling rate, wherein obtaining the second set of head poses comprises: generating, via an inertial measurement unit (IMU) associated with the user device, the second set of head poses at a second sampling rate.

[0247] Aspect 42 is the method of aspect 41, wherein the first sampling rate is less than the second sampling rate.

[0248] Aspect 43 is the method of any of aspects 29 to 42, wherein the first set of eye gazes comprises a third set of eye gaze time series data, wherein the second set of head poses comprises a fourth set of head pose time series data.

[0249] Aspect 44 is the method of any of aspects 29 to 43, wherein the first set of eye gazes comprises first translation information and first orientation information of at least one eye of a user associated with the user device, wherein the second set of head poses comprises second translation information and second orientation information of a head of the user associated with the user device.

[0250] Aspect 45 is the method of any of aspects 29 to 44, wherein outputting the third indication of the determined third set of predicted eye gazes comprises: transmitting the third indication of the determined third set of predicted eye gazes; receiving a fourth indication of a fourth set of rendered frames based on a transmission of the third indication of the determined third set of predicted eye gazes; and outputting a fifth indication of the fourth set of rendered frames.

[0251] Aspect 46 is the method of aspect 45, wherein outputting the fifth indication of the fourth set of rendered frames comprises: storing the fourth set of rendered frames in at least one of a memory, a buffer, or a cache.

[0252] Aspect 47 is the method of aspect 45, further comprising: decoding the fourth set of rendered frames; and warping the decoded fourth set of rendered frames based on a latest available head pose of a user associated with the user device, wherein outputting the fifth indication of the fourth set of rendered frames comprises: outputting a sixth indication of a fifth set of warped, decoded rendered frames.

[0253] Aspect 48 is the method of aspect 45, wherein each of the fourth set of rendered frames comprises a first region and a second region, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds to at least one of the third set of predicted eye gazes.

[0254] Aspect 49 is the method of any of aspects 29 to 48, wherein obtaining the first indication of the first set of eye gazes associated with the user device comprises: receiving the first indication of the first set of eye gazes from the user device, wherein obtaining the second indication of the second set of head poses associated with the user device comprises: receiving the second indication of the second set of head poses from the user device.

[0255] Aspect 50 is the method of aspect 49, wherein outputting the third indication of the determined third set of predicted eye gazes comprises: outputting the third indication to a Tenderer, further comprising: rendering, via the Tenderer, a fourth set of frames based on the determined third set of predicted eye gazes; and transmitting the rendered fourth set of frames to the user device.

[0256] Aspect 51 is the method of aspect 50, wherein rendering the fourth set of frames based on the determined third set of predicted eye gazes comprises: rendering a first region and a second region of each of the fourth set of frames, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds with at least one of the determined third set of predicted eye gazes.

[0257] Aspect 52 is the method of aspect 50, wherein transmitting the rendered fourth set of frames to the user device comprises: encoding the rendered fourth set of frames; and transmitting the encoded rendered fourth set of frames.

[0258] Aspect 53 is a method for graphics processing, comprising: obtaining an indication of an eye gaze associated with a user device; obtaining an indication of a head pose associated with the user device; determining a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose; and outputting an indication of the determined set of predicted eye gazes.

[0259] Aspect 54 is the method of aspect 53, further comprising: correlating the head pose with the eye gaze.

[0260] Aspect 55 is the method of aspect 54, wherein correlating the head pose with the eye gaze comprises: correlating the head pose with the eye gaze based on a first time indicator associated with the eye gaze and a second time indicator associated with the head pose.

[0261] Aspect 56 is the method of any of aspects 53 to 55, wherein the indication of the eye gaze comprises eye gaze time series data, wherein the indication of the head pose comprises head pose time series data.

[0262] Aspect 57 is the method of any of aspects 53 to 56, further comprising: obtaining an indication of a second eye gaze associated with the user device; obtaining an indication of a set of blinks associated with the user device; and determining that the second eye gaze is associated with at least one of the set of blinks, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises: refraining from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of blinks.

[0263] Aspect 58 is the method of any of aspects 53 to 57, further comprising: obtaining an indication of a second eye gaze associated with the user device; obtaining an indication of a set of saccades associated with the user device; and determining that the second eye gaze is associated with at least one of the set of saccades, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises: refraining from determining the set of predicted eye gazes based on the second eye gaze in responseto a determination that the second eye gaze is associated with at least one of the set of saccades.

[0264] Aspect 59 is the method of any of aspects 53 to 58, further comprising: converting the eye gaze into a six dimensional (6D) representation of the eye gaze, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises: determining the set of predicted eye gazes based on the converted 6D representation of the eye gaze.

[0265] Aspect 60 is the method of any of aspects 53 to 59, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises: determining the set of predicted eye gazes based on a first set of fully-connected network (FCN) blocks and a set of gated recurrent unit (GRU) blocks.

[0266] Aspect 61 is the method of any of aspects 53 to 59, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises: determining the set of predicted eye gazes based on a predictor model that uses the eye gaze and the head pose as inputs.

[0267] Aspect 62 is the method of aspect 61, wherein the predictor model comprises at least one of a regression model, a time series forecaster model, or a neural network.

[0268] Aspect 63 is the method of any of aspects 53 to 62, further comprising: obtaining an indication of a second eye gaze associated with the user device; determining that the second eye gaze is associated with a user tracking a virtual object; determining a second set of predicted eye gazes based on the obtained indication of the second eye gaze with respect to a head of the user in response to a determination that the eye gaze is associated with the user tracking the virtual object; and outputting an indication of the determined second set of predicted eye gazes.

[0269] Aspect 64 is the method of aspect 63, further comprising: determining that the eye gaze is not associated with the user tracking the virtual object, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises: determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises in response to a determination that the eye gaze is not associated with the user tracking the virtual object.

[0270] Aspect 65 is the method of any of aspects 53 to 64, wherein obtaining the indication of the eye gaze associate with the user device comprises: generating, via an eye gaze camera associated with the user device, a set of eye gazes comprising the eye gaze at a first sampling rate, wherein obtaining the head pose comprises: generating, via an inertial measurement unit (IMU) associated with the user device, a set of head poses comprising the head pose at a second sampling rate.

[0271] Aspect 66 is the method of aspect 65, wherein the first sampling rate is less than the second sampling rate.

[0272] Aspect 67 is the method of any of aspects 53 to 66, wherein the indication of the eye gaze comprises first translation information and first orientation information of at least one eye of a user associated with the user device, wherein the indication of the head pose comprises second translation information and second orientation information of a head of the user associated with the user device.

[0273] Aspect 68 is the method of any of aspects 53 to 67, wherein outputting the indication of the determined set of predicted eye gazes comprises: transmitting the indication of the determined set of predicted eye gazes; receiving an indication of a set of rendered frames after a transmission of the indication of the determined set of predicted eye gazes; and outputting an indication of the set of rendered frames.

[0274] Aspect 69 is the method of aspect 68, wherein outputting the indication of the set of rendered frames comprises: storing the set of rendered frames in at least one of a memory, a buffer, or a cache.

[0275] Aspect 70 is the method of aspect 68, further comprising: decoding the set of rendered frames; and warping the decoded set of rendered frames based on a latest available head pose of a user associated with the user device, wherein outputting the indication of the set of rendered frames comprises: outputting an indication of a set of warped, decoded rendered frames.

[0276] Aspect 71 is the method of aspect 68, wherein each of the set of rendered frames comprises a first region and a second region, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds to at least one of the set of predicted eye gazes.

[0277] Aspect 72 is the method of any of aspects 53 to 71, wherein obtaining the indication of the eye gaze associated with the user device comprises: receiving the indication ofthe eye gaze from the user device, wherein obtaining the indication of the head pose associated with the user device comprises: receiving the indication of the head pose from the user device.

[0278] Aspect 73 is the method of aspect 72, wherein outputting the indication of the determined set of predicted eye gazes comprises: outputting the indication to a Tenderer, wherein the method further comprises: rendering, via the Tenderer, a set of frames based on the determined set of predicted eye gazes; and transmitting the rendered set of frames to the user device.

[0279] Aspect 74 is the method of aspect 73, wherein rendering the set of frames based on the determined set of predicted eye gazes comprises: rendering a first region and a second region of each of the set of frames, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds with at least one of the determined set of predicted eye gazes.

[0280] Aspect 75 is the method of aspect 74, wherein transmitting the rendered set of frames to the user device comprises: encoding the rendered set of frames; and transmitting the encoded rendered set of frames.

[0281] Aspect 76 is an apparatus for graphics processing including at least one processor coupled to a memory and configured to implement a method as in any of aspects 1- 75.

[0282] Aspect 77 may be combined with aspect 76 and includes that the apparatus is a wireless communication device.

[0283] Aspect 78 is an apparatus for graphics processing including means for implementing a method as in any of aspects 1-75.

[0284] Aspect 79 is a computer-readable medium (e.g., a non-transitory computer-readable medium) storing computer executable code, the code when executed by at least one processor causes the at least one processor to implement a method as in any of aspects 1-75.

[0285] Various aspects have been described herein. These and other aspects are within the scope of the following claims.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. An apparatus for graphics processing, comprising: a memory; and a processor coupled to the memory and, based at least in part on information stored in the memory, the processor is configured to: obtain an indication of an eye gaze associated with a user device; obtain an indication of a head pose associated with the user device; determine a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose; and output an indication of the determined set of predicted eye gazes.

2. The apparatus of claim 1, wherein the processor is further configured to: correlate the head pose with the eye gaze.

3. The apparatus of claim 2, wherein, to correlate the head pose with the eye gaze, the processor is configured to: correlate the head pose with the eye gaze based on a first time indicator associated with the eye gaze and a second time indicator associated with the head pose.

4. The apparatus of claim 1, wherein the indication of the eye gaze comprises eye gaze time series data, wherein the indication of the head pose comprises head pose time series data.

5. The apparatus of claim 1, wherein the processor is further configured to: obtain an indication of a second eye gaze associated with the user device; obtain an indication of a set of blinks associated with the user device; and determine that the second eye gaze is associated with at least one of the set of blinks, wherein, to determine the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose, the processor is configured to:refrain from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of blinks.

6. The apparatus of claim 1, wherein the processor is further configured to: obtain an indication of a second eye gaze associated with the user device; obtain an indication of a set of saccades associated with the user device; and determine that the second eye gaze is associated with at least one of the set of saccades, wherein, to determine the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose, the processor is configured to: refrain from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of saccades.

7. The apparatus of claim 1, wherein the processor is further configured to: convert the eye gaze into a six dimensional (6D) representation of the eye gaze, wherein, to determine the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose, the processor is configured to: determine the set of predicted eye gazes based on the converted 6D representation of the eye gaze.

8. The apparatus of claim 1, wherein, to determine the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose, the processor is configured to: determine the set of predicted eye gazes based on a first set of fully-connected network (FCN) blocks and a set of gated recurrent unit (GRU) blocks.

9. The apparatus of claim 1, wherein, to determine the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose, the processor is configured to: determine the set of predicted eye gazes based on a predictor model that uses the eye gaze and the head pose as inputs.

10. The apparatus of claim 9, wherein the predictor model comprises at least one of a regression model, a time series forecaster model, or a neural network.

11. The apparatus of claim 1, wherein the processor is further configured to: obtain an indication of a second eye gaze associated with the user device; determine that the second eye gaze is associated with a user tracking a virtual obj ect; determine a second set of predicted eye gazes based on the obtained indication of the second eye gaze with respect to a head of the user in response to a determination that the eye gaze is associated with the user tracking the virtual object; and output an indication of the determined second set of predicted eye gazes.

12. The apparatus of claim 11, wherein the processor is further configured to: determine that the eye gaze is not associated with the user tracking the virtual object, wherein, to determine the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose, the processor is configured to: determine the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose in response to a determination that the eye gaze is not associated with the user tracking the virtual object.

13. The apparatus of claim 1, wherein, to obtain the indication of the eye gaze associate with the user device, the processor is configured to: generate, via an eye gaze camera associated with the user device, a set of eye gazes comprising the eye gaze at a first sampling rate, wherein, to obtain the head pose, the processor is configured to:generate, via an inertial measurement unit (IMU) associated with the user device, a set of head poses comprising the head pose at a second sampling rate.

14. The apparatus of claim 13, wherein the first sampling rate is less than the second sampling rate.

15. The apparatus of claim 1, wherein the indication of the eye gaze comprises first translation information and first orientation information of at least one eye of a user associated with the user device, wherein the indication of the head pose comprises second translation information and second orientation information of a head of the user associated with the user device.

16. The apparatus of claim 1, wherein, to output the indication of the determined set of predicted eye gazes, the processor is configured to: transmit the indication of the determined set of predicted eye gazes; receive an indication of a set of rendered frames after a transmission of the indication of the determined set of predicted eye gazes; and output an indication of the set of rendered frames.

17. The apparatus of claim 16, wherein the apparatus is a wireless communication device comprising at least one of a transceiver or an antenna coupled to the processor, wherein, to transmit the indication of the determined set of predicted eye gazes, the processor is configured to: transmit the indication of the determined set of predicted eye gazes via at least one of the transceiver or the antenna, wherein, to receive the indication of the set of rendered frames after the transmission of the indication of the determined set of predicted eye gazes, the processor is configured to: receive the indication of the set of rendered frames via at least one of the transceiver or the antenna.

18. The apparatus of claim 16, wherein, to output the indication of the set of rendered frames, the processor is configured to: store the set of rendered frames in at least one of a second memory, a buffer, or a cache.

19. The apparatus of claim 16, wherein the processor is further configured to: decode the set of rendered frames; and warp the decoded set of rendered frames based on a latest available head pose of a user associated with the user device, wherein, to output the indication of the set of rendered frames, the processor is configured to: output an indication of a set of warped, decoded rendered frames.

20. The apparatus of claim 16, wherein each of the set of rendered frames comprises a first region and a second region, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds to at least one of the set of predicted eye gazes.

21. The apparatus of claim 1, wherein, to obtain the indication of the eye gaze associated with the user device, the processor is configured to: receive the indication of the eye gaze from the user device, wherein, to obtain the indication of the head pose associated with the user device, the processor is configured to: receive the indication of the head pose from the user device.

22. The apparatus of claim 21, wherein, to output the indication of the determined set of predicted eye gazes, the processor is configured to: output the indication to a Tenderer, wherein the processor is further configured to: render, via the Tenderer, a set of frames based on the determined set of predicted eye gazes; and transmit the rendered set of frames to the user device.

23. The apparatus of claim 22, wherein the apparatus is a wireless communication device comprising at least one of a transceiver or an antenna coupled to the processor, wherein, to transmit the rendered set of frames to the user device, the processor is configured to: transmit the rendered set of frames to the user device via at least one of the transceiver or the antenna.

24. The apparatus of claim 22, wherein, to render the set of frames based on the determined set of predicted eye gazes, the processor is configured to: render a first region and a second region of each of the set of frames, wherein the first region is associated with a first quality level and the second region is associated with a second quality level, wherein the first quality level is greater than the second quality level, and wherein the first region corresponds with at least one of the determined set of predicted eye gazes.

25. The apparatus of claim 24, wherein, to transmit the rendered set of frames to the user device, the processor is configured to: encode the rendered set of frames; and transmit the encoded rendered set of frames.

26. The apparatus of claim 1, wherein the apparatus comprises a wireless communication device.

27. A method for graphics processing, comprising: obtaining an indication of an eye gaze associated with a user device; obtaining an indication of a head pose associated with the user device; determining a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose; and outputting an indication of the determined set of predicted eye gazes.

28. The method of claim 27, further comprising: obtaining an indication of a second eye gaze associated with the user device; obtaining an indication of a set of blinks associated with the user device; anddetermining that the second eye gaze is associated with at least one of the set of blinks, wherein determining the set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose comprises: refraining from determining the set of predicted eye gazes based on the second eye gaze in response to a determination that the second eye gaze is associated with at least one of the set of blinks.

29. The method of claim 27, further comprising: obtaining an indication of a second eye gaze associated with the user device; determining that the second eye gaze is associated with a user tracking a virtual object; determining a second set of predicted eye gazes based on the obtained indication of the second eye gaze with respect to a head of the user in response to a determination that the eye gaze is associated with the user tracking the virtual object; and outputting an indication of the determined second set of predicted eye gazes.

30. A computer-readable medium storing computer executable code, the code when executed by a processor, causes the processor to: obtain an indication of an eye gaze associated with a user device; obtain an indication of a head pose associated with the user device; determine a set of predicted eye gazes based on the obtained indication of the eye gaze and the obtained indication of the head pose; and output an indication of the determined set of predicted eye gazes.

Images