Light estimation for three-dimensional (3D) rendered objects

Abstract

A method for applying lighting conditions to a virtual object in an augmented reality (AR) device is described. In one aspect, the method includes generating an image using a camera of the mobile device; acquiring a virtual object corresponding to the object in the image; identifying illumination parameters of the virtual object based on a machine learning model pre-trained with a pairing dataset, the pairing dataset comprising synthesis source data and synthesis target data, the synthesis source data comprising an environmental map and a 3D scan of items depicted in the environmental map, the synthesis target data comprising a synthesis sphere image rendered in the same environmental map; the illumination parameter is applied to the virtual object, and the colored virtual object is displayed as a layer of the image in a display of the mobile device.

Description

[0001] Priority claim

[0002] This application claims the benefit of priority to U.S. application serial number 17 / 846,918, filed on June 22, 2022, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The subject matter disclosed herein generally relates to augmented reality (AR) systems. Specifically, the present disclosure proposes a method for light estimation of three-dimensional (3D) rendered virtual objects in an AR system. Background Art

[0004] Augmented reality (AR) enables a user to observe a scene while seeing related virtual content that can be aligned with items, images, objects, or environments in the field of view of the AR device. Therefore, the AR device mixes the rendered virtual content into the captured physical environment / scene (e.g., captured image) as much as possible to provide a more realistic experience to the user. However, some rendered virtual content appears out of context or unreal because the texture / brightness of the virtual content is inconsistent with the lighting conditions of the physical scene. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] In the drawings, which are not necessarily drawn to scale, the same reference numerals may describe similar components in different views. To easily identify the discussion of any particular element or action, one or more of the highest digits in the reference numerals refer to the figure number in which the element is first introduced. Some non-limiting examples are shown in the figures of the accompanying drawings, in which:

[0006] Figure 1 is a block diagram illustrating an environment for operating an AR device according to one example embodiment.

[0007] Figure 2 is a block diagram illustrating an AR device according to an example embodiment.

[0008] Figure 3 The training and use of a machine learning program according to some example embodiments is shown.

[0009] Figure 4 is a block diagram illustrating a light estimation system according to one example implementation.

[0010] Figure 5 is a block diagram illustrating a supervised training procedure according to an example implementation.

[0011] Figure 6 A method for training a machine learning model according to an example implementation is shown.

[0012] Figure 7A method for applying shades to virtual objects using a machine learning model is shown according to an example implementation.

[0013] Figure 8 Examples of virtual objects that are independent of lighting conditions and virtual objects that are dependent on lighting conditions are shown according to an example embodiment.

[0014] Fig. 9 An example of source data and target data according to one example implementation is shown.

[0015] Fig.10 Predicted lighting conditions for a virtual object are shown according to an example implementation.

[0016] Fig.11 An example of a shaded virtual object is shown according to one embodiment.

[0017] Fig.12 An example of a shaded virtual object is shown according to one embodiment.

[0018] Fig.13 is a block diagram illustrating a software architecture in which the present disclosure may be implemented according to an example implementation.

[0019] Fig.14 is a diagrammatic representation of a machine in the form of a computer system according to an example embodiment within which a set of instructions may be executed, causing the machine to perform any one or more of the methodologies discussed herein. DETAILED DESCRIPTION

[0020] The following description describes the system, method, technology, instruction sequence and computer program product of the example implementation of this theme. In the following description, for the purpose of illustration, many specific details are set forth to provide an understanding of the various implementations of this theme. However, it will be apparent to those skilled in the art that the implementation of this theme can be put into practice without some or other details in these specific details. Examples only represent possible variations. Unless otherwise clearly stated, structures (e.g., structural components, such as modules) are optional and can be combined or subdivided, and operations (e.g., in processes, algorithms or other functions) can be changed in order or combined or subdivided.

[0021] The term "augmented reality" (AR) is used herein to refer to an interactive experience of a real-world environment in which physical objects residing in the real world are "augmented" or enhanced by computer-generated digital content (also referred to as virtual content or synthetic content). AR may also refer to a system that enables the combination of the real and virtual worlds, real-time interaction, and 3D registration of virtual and real objects. Users of AR systems perceive virtual content that appears to be attached / anchored / interacting with physical objects in the real world. The term "AR application" is used herein to refer to a computer-operated application that enables an AR experience.

[0022] The term "AR device" is used herein to refer to a computing device that operates an AR application. The AR device allows a user to access information, for example in the form of virtual content rendered in a display of the AR device (also referred to as a display device). The rendering of the virtual content can be based on the positioning of the display device relative to physical objects or relative to a reference frame (external to the display device) so that the virtual content appears correctly in the display.

[0023] Mix as many rendered AR objects as possible into the environment to provide a more realistic experience to the user of the AR device. For example, an AR device enables a user to visualize a product (e.g., sunglasses, shoes, watches) on their body or in a physical environment (e.g., a car, furniture). Typical rendered virtual content does not take into account the lighting conditions of the physical environment. In a situation where an AR device renders a pair of sunglasses on the user's face without considering the lighting conditions in the original scene, the user will feel the "fakeness" of the rendered sunglasses. To the user, the rendered virtual objects appear to "jump out" of the screen, looking unnatural and unreal.

[0024] This application describes a machine learning system for more realistic rendering by estimating lighting conditions from camera textures. In one example, the system collects a set of HDRI (High Dynamic Range Imaging) environment maps and a set of 3D scans of real people. The system uses this data to render a paired dataset, where the source image is a person rendered within an HDRI environment map, and the target image is a sphere rendered in the same HDRI environment map.

[0025] During model training, a cropped face from a source image is fed into a neural network that predicts light parameters (e.g., a spherical Gaussian that approximates the lighting conditions, ambient light). A differential renderer uses the predicted parameters to render a new sphere (e.g., a predicted sphere image). The system then compares the new sphere to the corresponding target sphere. During model inference, the machine learning system provides the predicted parameters to a rendering engine (e.g., a physically based rendering (PBR) shader) to apply the estimated lighting conditions to any rendered virtual objects in the scene.

[0026] In one example implementation, the present application describes a method for estimating lighting conditions for rendering virtual objects for an AR system. In one aspect, the method includes: generating an image using a camera of a mobile device; acquiring a virtual object corresponding to an object in the image; identifying shading parameters of the virtual object based on a machine learning model pre-trained with a paired data set, the paired data set including synthetic source data and synthetic target data, the synthetic source data including an environment map and a 3D scan of an item depicted in the environment map, the synthetic target data including a synthetic sphere image rendered in the same environment map; applying the shading parameters to the virtual object; and displaying the shaded virtual object as a layer of the image in a display of the mobile device.

[0027] Therefore, one or more of the methods described herein help solve the technical problem of rendering virtual objects in an AR device.The presently described method provides an improvement to the operation of realistic rendering for an AR device by estimating lighting conditions from camera textures.

[0028] Figure 1 1 is a network diagram illustrating an environment 100 suitable for operating an AR device 106 according to some example implementations. The environment 100 includes a user 104 and an AR device 106. The user 104 operates the AR device 106. For example, the user 104 is using the AR device 106 to capture a self-portrait image (also referred to as a selfie) of the user's face. The user 104 can be a human user (e.g., a human), a machine user (e.g., a computer configured by a software program to interact with the AR device 106), or any suitable combination thereof (e.g., a human assisted by a machine or a machine supervised by a human).

[0029] The AR device 106 may be a computing device with a display, such as a smartphone, a tablet, or a wearable computing device (e.g., a watch or glasses). The computing device may be handheld or may be removably mounted to the head of the user 104. In one example, the display includes a screen that displays images captured by a camera of the AR device 106. In another example, the display of the device may be transparent (e.g., in the lenses of wearable computing glasses), which enables the user 104 to view virtual content presented on the display while also being able to view real-world objects in the user's 104 line of sight through the display. In other examples, the display may be opaque, partially transparent, or partially opaque. In other examples, the display may be worn by the user 104 to cover a portion of the user's 104 field of view.

[0030] The AR device 106 includes an augmented reality system 110 that generates lighting-dependent (also referred to as "shaded") virtual content based on images detected with a camera of the AR device 106. For example, the user 104 may direct the camera of the AR device 106 to capture an image of the user's 104 face or a physical object (not shown) in a scene of the real-world environment 102. The augmented reality system 110 generates lighting-dependent virtual content (e.g., tinted virtual sunglasses) corresponding to an object identified in the image (e.g., the user's 104 face) based on existing lighting conditions in the scene of the real-world environment 102. An example of a scene includes a portion of the real-world environment 102 captured by the camera of the AR device 106.

[0031] The AR device 106 presents the tinted virtual content (e.g., tinted virtual sunglasses) in a display of the AR device 106. In another example, the augmented reality system 110 renders the virtual content, applies tinting based on the existing lighting conditions of the scene, and presents the tinted virtual content in the display of the AR device 106 relative to a reference frame (external to the display device) so that the virtual content appears correctly in the display. In other words, the tinted virtual content (e.g., virtual sunglasses) appears anchored to the face to the user 104.

[0032] In one example implementation, the augmented reality system 110 includes a light estimation system 108 and an AR application 112. The AR application 112 uses computer vision to detect and recognize physical environments (e.g., real-world environment 102), items (e.g., face of user 104, shoes, body parts of the user such as wrists) depicted in images captured by a camera of the AR device 106. The AR application 112 retrieves virtual content (e.g., a 3D object model of sunglasses) based on the items / physical objects or scenes identified in the real-world environment 102. The AR application 112 renders the virtual objects in the display 204. In another example, the AR application 112 obtains estimated light conditions (e.g., lighting / shading parameters) from the light estimation system 108 and applies the predicted lighting parameters to a physically based renderer (also referred to as PBR), which applies the estimated light conditions to the virtual objects (e.g., sunglasses). The AR application 112 displays the tinted virtual objects (e.g., tinted sunglasses) as an overlay on the face of the user 104.

[0033] The light estimation system 108 identifies predicted lighting parameters (e.g., shading parameters, ambient light) for a virtual object / item (e.g., sunglasses) based on the lighting conditions in the image of the face of the user 104. In another example, the light estimation system 108 obtains a virtual object model from the AR application 112 and applies the predicted lighting parameters (e.g., estimated ambient light conditions) to the virtual object (e.g., sunglasses) using a renderer to generate a more realistic virtual object consistent with the lighting conditions in the scene captured by the camera of the AR device 106. The light estimation system 108 includes a machine learning model trained using supervised training with synthetic data. Figure 4 Light estimation system 108 is described in more detail.

[0034] In another example embodiment, the augmented reality system 110 includes other applications, such as a 6DOF tracking system or a depth sensing system.

[0035] Figure 1 Any of the machines, databases, or devices shown in the can be implemented in a general-purpose computer that is modified (e.g., configured or programmed) by software to be a special-purpose computer to perform one or more of the functions described herein for the machine, database, or device. As used herein, a "database" is a data storage resource and can store data structured as a text file, table, spreadsheet, relational database (e.g., object-relational database), triple store, hierarchical data storage, or any suitable combination thereof. In addition, Figure 1 Any two or more of the machines, databases, or devices shown in may be combined into a single machine, and the functionality described herein for any single machine, database, or device may be divided among multiple machines, databases, or devices.

[0036] The AR device 106 may operate via a computer network. A computer network may be any network that enables communication between or among machines, databases, and devices. Thus, a computer network may be a wired network, a wireless network (e.g., a mobile or cellular network), or any suitable combination thereof. A computer network may include one or more parts that constitute a private network, a public network (e.g., the Internet), or any suitable combination thereof.

[0037] Figure 2 2 is a block diagram illustrating modules (e.g., components) of an AR device 106 according to some example implementations. The AR device 106 includes a sensor 202, a display 204, a display controller 218, an image processing unit 216, a processor 208, and a storage device 206. Examples of the AR device 106 include a wearable computing device, a mobile computing device, a navigation device, a smart phone, and the like.

[0038] Sensors 202 include, for example, optical sensors 210 (e.g., cameras such as color cameras, thermal cameras, depth sensors, and one or more grayscale, global / rolling shutter tracking cameras) and inertial sensors 212 (e.g., gyroscopes, accelerometers). Other examples of sensors 202 include proximity sensors or position sensors (e.g., near field communication, GPS, Bluetooth, Wi-Fi), audio sensors (e.g., microphones), or any suitable combination thereof. It should be noted that the sensors 202 described herein are for illustrative purposes, and thus the sensors 202 are not limited to the sensors described above.

[0039] The processor 208 implements and operates the AR application 112 and the light estimation system 108. The AR application 112 uses computer vision to detect and identify physical items / objects (e.g., the face of the user 104) or the physical environment (e.g., the real world environment 102). The AR application 112 retrieves virtual content (e.g., a 3D object model) based on the identified physical items (e.g., the face of the user 104) or the physical environment. The AR application 112 renders the virtual object in the display 204 so that the virtual object appears to be anchored to the physical environment or physical item. In an example embodiment, the AR application 112 includes a local rendering engine that generates a visualization of the virtual content superimposed on (e.g., superimposed on or otherwise displayed together with) an image of the physical item captured by the optical sensor 210. The visualization of the virtual content can be manipulated by adjusting the position of the physical item relative to the optical sensor 210 (e.g., its physical position, orientation, or both). Similarly, the visualization of the virtual content can be manipulated by adjusting the posture of the AR device 106 relative to the physical item.

[0040] In one example, the AR application 112 uses sensor data from an IMU sensor (e.g., inertial sensor 212) to obtain rotational motion data of the AR device 106. The processor 208 captures rotational and translational motion data of the AR device 106. The processor 208 uses image data from the optical sensor 210 and corresponding inertial data from the inertial sensor 212 to track the position and pose of the AR device 106 relative to a reference frame (e.g., real-world environment 102, face of user 104).

[0041] The light estimation system 108 identifies predicted lighting parameters (e.g., shading parameters) for a virtual object / item (e.g., sunglasses) based on the lighting conditions in the image captured by the optical sensor 210. In another example, the light estimation system 108 applies the predicted lighting parameters (e.g., estimated light conditions) to the texture of the virtual object (e.g., sunglasses) using a rendering engine (not shown) at the graphics processing unit 216 to generate a more realistic virtual object that is consistent with the lighting conditions in the scene captured by the optical sensor 210 of the AR device 106. The light estimation system 108 includes a machine learning model trained using synthetic data of an HDRI environment map and a 3D scan of a person. Figure 4 Light estimation system 108 is described in more detail.

[0042] The graphics processing unit 216 includes a rendering engine (not shown) configured to render frames / textures / shading of 3D models of virtual objects based on the virtual content provided by the AR application, the pose of the AR device 106, and the shading parameters from the light estimation system 108. In other words, the graphics processing unit 216 uses the three-dimensional pose of the AR device 106 to generate frames of rendered virtual content to be presented on the display 204. For example, the graphics processing unit 216 uses the three-dimensional pose to render frames of virtual content so that the virtual content is presented at a certain orientation and position in the display 204 to appropriately enhance the user's sense of reality. As an example, the graphics processing unit 216 can use the three-dimensional pose data to render frames of virtual content so that when the virtual content is presented on the display 204, the virtual content overlaps with physical objects in the user's real-world environment 102. The graphics processing unit 216 generates an updated frame of the rendered virtual content based on the updated three-dimensional pose of the graphics processing unit 216, and the updated three-dimensional pose reflects the change in the position and orientation of the user relative to the physical objects in the user's real-world environment 102.

[0043] The graphics processing unit 216 transmits the rendered frames to the display controller 218. The display controller 218 is positioned as an intermediary between the display controller 218 and the display 204, receiving image data (eg, rendered frames) from the graphics processing unit 216 and providing the rendered frames to the display 204.

[0044] The display 204 comprises a screen or monitor configured to display images generated by the processor 208. In one example implementation, the display 204 may be transparent or translucent so that the user 104 can see through the display 204 (in an AR use case). In another example implementation, the display 204 covers the eyes of the user 104 and occludes the entire field of view of the user 104 (in a VR use case). In another example, the display 204 comprises a touch screen display configured to receive user input via contact on the touch screen display.

[0045] Storage device 206 stores virtual content 214. Virtual content 214 includes, for example, a database of machine learning models of lighting conditions, visual references (e.g., images of physical objects), and corresponding experiences (e.g., three-dimensional virtual object models). Other augmented data that may be stored in storage device 206 include augmented reality content items (e.g., corresponding to application lenses or augmented reality experiences). Augmented reality content items may be real-time special effects and sounds that may be added to images or videos.

[0046] As described above, augmented data includes augmented reality content items, overlays, image transformations, AR images, and similar terms that refer to modifications that can be applied to image data (e.g., video or images). This includes real-time modifications that modify images when they are captured using the device sensor (e.g., one or more cameras) of the AR device 106 and then display the modified images on the screen of the AR device 106. This also includes modifications to stored content (e.g., video clips in a gallery that can be modified). For example, in an AR device 106 that accesses multiple augmented reality content items, a user can use a single video clip with multiple augmented reality content items to see how different augmented reality content items will modify the stored clips. For example, by selecting different augmented reality content items for the same content, multiple augmented reality content items that apply different pseudo-random motion models can be applied to the same content. Similarly, real-time video capture can be used with the modifications shown to show how the video image currently being captured by the sensor of the AR device 106 will modify the captured data. Such data may be displayed only on the screen and not stored in memory, or content captured by the device sensors may be recorded and stored in memory with or without modification (or both). In some systems, a preview function may simultaneously display how different augmented reality content items will look in different windows of the display. For example, this may enable viewing multiple windows with different pseudo-random animations on the display at the same time.

[0047] Thus, data and various systems using augmented reality content items or other such transformation systems using the data to modify content may involve: detecting objects (e.g., faces, hands, bodies, cats, dogs, surfaces, objects, etc.) in video frames, tracking such objects as they leave, enter, and move around the field of view, and modifying or transforming such objects as they are tracked. In various examples, different methods for implementing such transformations may be used. Some examples may involve: generating a three-dimensional mesh model of one or more objects, and using transformations and animated textures of the models within the video to implement the transformations. In other examples, tracking of points on an object may be used to place an image or texture (which may be two-dimensional or three-dimensional) at the tracked location. In other examples, neural network analysis of video frames may be used to place an image, model, or texture in content (e.g., an image or video frame). Thus, augmented reality content items involve both images, models, and textures used to create transformations in content, and additional modeling and analysis information required to implement such transformations using object detection, tracking, and placement.

[0048] Real-time video processing can be performed using any kind of video data (e.g., video streams, video files, etc.) stored in the memory of any kind of computerized system. For example, a user can load a video file and save it in the memory of the device, or a video stream can be generated using the device's sensors. In addition, computer animation models can be used to process any object, such as human faces and parts of the human body, animals, or non-living things (e.g., chairs, cars, or other objects).

[0049] In some examples, in the case of selecting a specific modification together with the content to be transformed, the element to be transformed is identified by a computing device, and then if the element to be transformed is present in a frame of a video, the element to be transformed is detected and tracked. The elements of the object are modified according to the modification request, so the frame of the video stream is transformed. For different kinds of transformations, the transformation of the frame of the video stream can be performed by different methods. For example, for the transformation of the frame that mainly involves changing the form of the elements of the object, the feature points of each element of the object are calculated (for example, using an active shape model (ASM) or other known methods). Then, a grid based on the feature points is generated for each element of at least one element of the object. The grid is used for the subsequent stage of tracking the elements of the object in the video stream. During the tracking process, the grid mentioned for each element is aligned with the position of each element. Then, additional points are generated on the grid. A first set of first points is generated for each element based on the modification request, and a set of second points is generated for each element based on the set of first points and the modification request. Then, the frame of the video stream can be transformed by modifying the elements of the object based on the set of first points, the set of second points and the grid. In such a method, the background of the modified object can also be changed or distorted by tracking and modifying the background.

[0050] In some examples, the transformation of certain areas of an object using elements of the object can be performed by calculating feature points of each element of the object and generating a grid based on the calculated feature points. Points are generated on the grid, and then various areas based on these points are generated. Then, the elements of the object are tracked by aligning the area of ​​each element with the position of each element of at least one element, and the properties of the area can be modified based on the modification request, so the frame of the video stream is transformed. Depending on the specific modification request, the properties of the mentioned area can be transformed in different ways. Such modifications may involve: changing the color of the area; removing at least some parts of the area from the frame of the video stream; including one or more new objects in the area based on the modification request; and modifying or distorting the elements of the area or object. In various examples, any combination of such modifications or other similar modifications may be used. For certain models to be animated, some characteristic points may be selected as control points of the entire state space to be used to determine the options for model animation.

[0051] In some examples of computer animation models that use face detection to transform image data, faces are detected on an image using a specific face detection algorithm (e.g., Viola-Jones). An active shape model (ASM) algorithm is then applied to the face region of the image to detect facial feature reference points.

[0052] Other methods and algorithms suitable for face detection can be used. For example, in some examples, landmarks are used to locate features, which represent distinguishable points that are present in most of the images considered. For example, for facial landmarks, the location of the left eye pupil can be used. If the initial landmarks are not identifiable (for example, if the person has an eye mask), secondary landmarks can be used. Such a landmark identification process can be used for any such object. In some examples, the set of landmarks forms a shape. The shape can be represented as a vector using the coordinates of the points in the shape. One shape is aligned with another shape using a similarity transformation (allowing translation, scaling, and rotation) that minimizes the average Euclidean distance between the shape points. The mean shape (meanshape) is the average value of the aligned training shapes.

[0053] In some examples, a landmark search begins with an average shape that aligns with the location and size of a face as determined by a global face detector. Such a search then repeats the steps of suggesting tentative shapes by adjusting the positioning of shape points through template matching of the image texture around each point, and then conforming the tentative shapes to a global shape model until convergence occurs. In some systems, template matching alone is unreliable, and the shape model pools the results of weak template matches to form a stronger overall classifier. The entire search is repeated at each level of the image pyramid, from coarse to fine resolution.

[0054] The transformation system can capture an image or video stream on a client device (e.g., AR device 106) and perform complex image manipulations locally on the AR device 106 while maintaining an appropriate user experience, computation time, and power consumption. Complex image manipulations can include size and shape changes, emotion conversions (e.g., changing a face from a frown to a smile), state transitions (e.g., aging an object, reducing apparent age, changing gender), style transfer, application of graphical elements, and any other suitable image or video manipulations enabled by a convolutional neural network that has been configured to execute efficiently on the AR device 106.

[0055] In some examples, a computer animation model for transforming image data may be used by a system in which a user may capture an image or video stream of the user (e.g., a selfie) using an AR device 106 having a neural network operating as part of an AR application operating on the AR device 106. A transformation system operating within the AR application determines the presence of a face within the image or video stream and provides a modification icon associated with the computer animation model to transform the image data, or the computer animation model may exist in association with an interface described herein. The modification icon includes a change that may be a basis for modifying the user's face within the image or video stream as part of the modification operation. Once the modification icon is selected, the transformation system initiates a process of transforming the user's image to reflect the selected modification icon (e.g., generating a smiley face on the user). Once the image or video stream is captured and the specified modification is selected, the modified image or video stream may be presented in a graphical user interface displayed on the AR device 106. The transformation system may implement a complex convolutional neural network on a portion of the image or video stream to generate and apply the selected modification. That is, a user can capture an image or video stream, and once a modification icon has been selected, the modification results are presented to the user in real time or near real time. Furthermore, while the video stream is being captured, the modification can be persistent and the selected modification icon remains toggled. Machine-taught neural networks can be used to implement such modifications.

[0056] The graphical user interface presenting the modifications performed by the transformation system can provide additional interactive options to the user. Such options can be based on the interface used to initiate the selection of a particular computer animation model and content capture (e.g., initiated from a content creator user interface). In various examples, the modifications can be persistent after the initial selection of the modification icon. The user can turn the modification on or off by tapping or otherwise selecting the face being modified by the transformation system, and store it for later viewing or browsing to other areas of the imaging application. In the case of multiple faces being modified by the transformation system, the user can globally turn the modification on or off by tapping or selecting a single face modified and displayed within the graphical user interface. In some examples, each face in a group of multiple faces can be modified individually, or such modifications can be individually switched by tapping or selecting each face or a series of each face displayed within the graphical user interface.

[0057] Any one or more modules in the modules described herein can be implemented using hardware (e.g., a processor of a machine) or a combination of hardware and software. For example, any module described herein can configure a processor to perform the operation described herein for the module. In addition, any two or more modules in these modules can be combined into a single module, and the function described herein for a single module can be subdivided between multiple modules. In addition, according to various example embodiments, the modules described herein as implemented in a single machine, database, or device can be distributed across multiple machines, databases, or devices.

[0058] Figure 3 The training and use of a machine learning program 300 is shown according to some example embodiments. In some example embodiments, a machine learning program (MLP) (also referred to as a machine learning algorithm or tool) is used to perform operations associated with lighting condition parameters.

[0059] Machine learning is a field of study that gives computers the ability to learn without explicit programming. Machine learning explores the study and construction of algorithms (also referred to herein as tools) that can learn from existing data and make predictions about new data. Such machine learning tools operate by building models from example training data 304 to make data-driven predictions or decisions represented as outputs or evaluations (e.g., evaluations 312). Although example implementations are presented with respect to some machine learning tools, the principles presented herein can be applied to other machine learning tools.

[0060] In some example implementations, different machine learning tools may be used. For example, logistic regression (LR), naive Bayes, random forest (RF), neural network (NN), matrix factorization, and support vector machine (SVM) tools may be used to classify or score job postings.

[0061] Two common types of problems in machine learning are classification problems and regression problems. Classification problems (also known as category problems) aim to classify items into one of several category values, for example, classifying objects into types of fruit (e.g., apples or oranges). Regression algorithms aim to quantify some terms, for example, by providing values ​​as real numbers.

[0062] The machine learning algorithm uses features 302 to analyze the data to generate estimates 312. Each feature in features 302 is a separate measurable attribute of the observed phenomenon. The concept of a feature is related to the concept of an explanatory variable used in statistical techniques such as linear regression. Selecting features that are informative, discriminating, and independent is important for the effective operation of MLPs in pattern recognition, classification, and regression. Features can be of different types, such as numeric features, strings, and graphics.

[0063] In an example implementation, features 302 may be of different types and may include, by way of example only, one or more of content 314 , concepts 316 , attributes 318 , historical data 320 , and / or user data 322 .

[0064] The machine learning algorithm uses the training data 304 to find correlations between the identified features 302 that influence the outcome or evaluation 312. In some example embodiments, the training data 304 includes labeled data, which is known data for one or more identified features 302 and one or more outcomes (e.g., detected lighting patterns or parameters).

[0065] Using the training data 304 and the identified features 302, the machine learning tool is trained at machine learning program training 306. The machine learning tool evaluates the value of the feature 302 when the feature 302 is associated with the training data 304. The result of the training is a trained machine learning program 310.

[0066] When performing an evaluation using the trained machine learning program 310, new data 308 is provided as input to the trained machine learning program 310, and the trained machine learning program 310 generates an evaluation 312 as output.

[0067] Figure 4 4 is a block diagram illustrating a light estimation system 108 according to an example implementation. The light estimation system 108 includes a supervised training program 406, a trained machine learning model 404, and a shader application 402.

[0068] The supervised training program 406 uses training data 408 (e.g., synthetic data) to train a machine learning model (e.g., trained machine learning model 404). In one example, the training data 408 includes a set of HDRI environment maps and a set of 3D scans of real people. The supervised training program 406 uses the training data 408 to render a paired data set, wherein the source image is an object (e.g., a person) rendered within an HDRI environment map, and the target image is a sphere rendered in the same HDRI environment map. In one example implementation, during model training, the training data 408 (e.g., a cropped face from a 3D scan of a real person from the source image) is fed into the neural network of the supervised training program 406, and lighting parameters (e.g., spherical Gaussian, ambient light) are predicted. The supervised training program 406 uses these predicted parameters to render a new predicted sphere image, and the rendered new predicted sphere image is compared to the corresponding target synthetic sphere image using an L2 loss.

[0069] The trained machine learning model 404 can be used to generate predicted lighting parameters based on an image of an object used during training (e.g., the face of the user 104). For example, the AR application 112 captures a self-portrait of the user 104. The AR application 112 identifies virtual sunglasses to be rendered on the face of the user 104. The AR application 112 provides the image data (the self-portrait image) and the virtual object model (e.g., a textured 3D model of the sunglasses) to the trained machine learning model 404. The light estimation system 108 applies the trained machine learning model 404 to generate predicted lighting parameters for the virtual object.

[0070] During model inference, the trained machine learning model 404 provides the predicted lighting parameters to the shader application 402 to apply the lighting parameters to the shading texture of the 3D model. The shader application 402 includes, for example, a PBR shader engine. The shader application 402 provides a shaded virtual object based on the AR application 112. The AR application 112 displays the shaded virtual object in the display 204.

[0071] Figure 5 is a block diagram illustrating a supervised training procedure 406 according to an example implementation. The supervised training procedure 406 includes a facial asset 502, an HDR environment map 504, a sphere asset 506, a renderer A 508, a renderer B 510, a synthesized facial image 512, a synthesized sphere image 514, a neural network 516, a spherical Gaussian 518, an ambient light 520, a differential renderer 522, an L2 loss 524, a predicted sphere image 526, and a lighting prediction module 528.

[0072] Training data 408 includes a facial asset 502, an HDR environment map 504, and a sphere asset 506. Examples of facial assets 502 include cropped 3D scans of real people. Examples of HDR environment maps 504 include HDRI maps taken from different locations. Examples of sphere assets 506 include synthetic sphere 3D models with configurable materials. Although an example of a face is given, the same techniques can be used for any object to correct the lighting of a virtual object rendered with the object.

[0073] Renderer A 508 includes a rendering engine that renders a composite facial image 512 based on the facial asset 502 and the HDR environment map 504. The composite facial image 512 is fed into a neural network 516 to predict a lighting prediction module 528 (e.g., parameters of a spherical Gaussian 518 and ambient light 520). A differential renderer 522 renders a predicted sphere image 526 based on the sphere asset 506 and the parameters of the spherical Gaussian 518 and ambient light 520.

[0074] Renderer B 510 includes a rendering engine that renders a synthesized sphere image 514 based on the HDR environment map 504 and the sphere asset 506. An L2 loss 524 is used to compare the synthesized sphere image 514 to a predicted sphere image 526. In another example, the result of the L2 loss 524 is used to train a neural network via back-propagation.

[0075] Figure 6 6 is a flow chart illustrating a method 600 according to an example implementation. The operations of the method 600 may be performed by the augmented reality system 110. In one example, the method 600 may operate with the light estimation system 108.

[0076] In block 602, the supervised training program 406 obtains a synthesized facial image (source data). In block 604, the supervised training program 406 generates predicted lighting parameters (e.g., spherical Gaussian, ambient light) using the NN. In block 606, the supervised training program 406 generates a predicted sphere image based on the predicted lighting parameters using a differential renderer. In block 608, the supervised training program 406 compares the predicted sphere image to the synthesized sphere image (target data).

[0077] It should be noted that other embodiments can use different sorting, additional or less operations and different naming or terminology to implement similar functions. In some embodiments, various operations can be performed in parallel with other operations in a synchronous or asynchronous manner. The operations described herein are selected to illustrate some operating principles in a simplified form.

[0078] Figure 7 7 is a flow chart illustrating a method 700 according to one example implementation. The operations of the method 700 may be performed by the augmented reality system 110, the light estimation system 108, or any combination thereof.

[0079] In block 702, the AR application 112 acquires an image. In block 704, the AR application 112 identifies a virtual object. In block 706, the light estimation system 108 generates shading parameters for the virtual object based on the image using the trained model. In block 708, the light estimation system 108 applies the shading parameters to the virtual object. In block 710, the AR application 112 displays the shaded virtual object as an overlay on the image.

[0080] It should be noted that other embodiments can use different sorting, additional or less operations and different naming or terminology to implement similar functions. In some embodiments, various operations can be performed in parallel with other operations in a synchronous or asynchronous manner. The operations described herein are selected to illustrate some operating principles in a simplified form.

[0081] Figure 8Examples of virtual objects that are independent of lighting conditions and virtual objects that are dependent on lighting conditions are shown according to an example implementation. Uncolored virtual object image 802 shows virtual sunglasses 806 colored with fixed lighting parameters, which appear "fake" with the image of the user. Colored virtual object image 804 shows virtual sunglasses 808 colored with predicted lighting parameters on the user's face. Virtual sunglasses 808 colored with predicted lighting parameters are colored based on the lighting conditions in the user's self-portrait image.

[0082] Fig. 9 An example of source data and target data according to an example implementation is shown. Examples of source data include source image 902 (e.g., a 3D facial scan of a person 908) and an HDRI environment map 906). Examples of target data include target image 904 (e.g., a synthetic 3D sphere rendered in the same HDRI environment map 906).

[0083] Fig.10 1004 is based on a comparison between the synthetic sphere image 1002 and the predicted sphere image 1006.

[0084] Fig.11 An example of a rendered virtual object is shown according to one embodiment. The self-portrait image 1106 depicts the user's face 1102 with rendered virtual content 1104 .

[0085] Fig.12 An example of a rendered virtual object is shown according to one embodiment. The example display 1202 depicts an image of a car 1204 with a rendered virtual object 1206 (captured by a camera of the AR device 106 ).

[0086] Fig.13 1300 is a block diagram illustrating a software architecture 1304 that can be installed on any one or more of the devices described herein. The software architecture 1304 is supported by hardware such as 1302 including a processor 1320, a memory 1326, and an I / O component 1338. In this example, the software architecture 1304 can be conceptualized as a stack of layers, where each layer provides specific functionality. The software architecture 1304 includes layers such as an operating system 1312, a library 1310, a framework 1308, and an application 1306. In operation, the application 1306 invokes an API call 1350 through the software stack and receives a message 1352 in response to the API call 1350.

[0087] The operating system 1312 manages hardware resources and provides public services. The operating system 1312 includes, for example, a kernel 1314, services 1316, and drivers 1322. The kernel 1314 acts as an abstraction layer between the hardware and other software layers. For example, the kernel 1314 provides memory management, processor management (e.g., scheduling), component management, networking and security settings, and other functions. Services 1316 can provide other public services to other software layers. Drivers 1322 are responsible for controlling or interfacing with the underlying hardware. For example, drivers 1322 may include display drivers, camera drivers, or Low-power drivers, Flash drivers, Serial communications drivers (e.g., Universal Serial Bus (USB) drivers), Drivers, audio drivers, power management drivers, etc.

[0088] The library 1310 provides a low-level common infrastructure used by the application 1306. The library 1310 may include 1318 (e.g., a C standard library) that provides functions such as memory allocation functions, string manipulation functions, mathematical functions, etc. In addition, the library 1310 may include 1324, such as a media library (e.g., a library for supporting the presentation and manipulation of various media formats, such as Moving Picture Experts Group-4 (MPEG4), Advanced Video Coding (H.264 or AVC), Moving Picture Experts Group Layer-3 (MP3), Advanced Audio Coding (AAC), Adaptive Multi-Rate (AMR) audio codec, Joint Photographic Experts Group (JPEG or JPG), or Portable Network Graphics (PNG)), a graphics library (e.g., an OpenGL framework for presenting in two dimensions (2D) and three dimensions (3D) in graphics content on a display), a database library (e.g., SQLite for providing various relational database functions), a web library (e.g., WebKit for providing web browsing functions), etc. The library 1310 may also include a variety of 1328 to provide many other APIs to the application 1306.

[0089] The framework 1308 provides a high-level common infrastructure used by the applications 1306. For example, the framework 1308 provides various graphical user interface (GUI) functions, high-level resource management, and high-level positioning services. The framework 1308 can provide a wide range of other APIs that can be used by the applications 1306, some of which can be specific to a particular operating system or platform.

[0090] In an example implementation, applications 1306 may include 1336, 1330, 1332, 1334, 1342, 1344, 1346, 1348, and a variety of other applications such as third-party applications 1340. Application 1306 is a program that executes the functions defined in the program. Various programming languages ​​may be used to create one or more of the applications 1306 structured in various ways, such as an object-oriented programming language (e.g., Objective-C, Java, or C++) or a procedural programming language (e.g., C language or assembly language). In a specific example, third-party applications 1340 (e.g., provided by an entity other than the vendor of a particular platform using ANDROID TM or IOS TM Software Development Kit (SDK) can be used to develop applications on platforms such as IOS TM ANDROID TM , Mobile software running on the mobile operating system of the phone or other mobile operating systems. In this example, third-party applications 1340 can activate API calls 1350 provided by the operating system 1312 to facilitate the functions described herein.

[0091] Fig.141400 in which instructions 1408 (e.g., software, programs, applications, applet, app, or other executable code) may be executed to cause the computer 1400 to perform any one or more of the methods discussed herein. For example, the instructions 1408 may cause the computer 1400 to perform any one or more of the methods described herein. The instructions 1408 convert a general purpose, unprogrammed computer 1400 into a specific computer 1400 that is programmed to perform the functions described and illustrated in the manner described. The computer 1400 may operate as a standalone device or may be coupled (e.g., networked) to other machines. In a networked deployment, the computer 1400 may operate as a server machine or a client machine in a server-client network environment, or as a peer machine in a peer-to-peer (or distributed) network environment. The computer 1400 may include, but is not limited to, a server computer, a client computer, a personal computer (PC), a tablet computer, a laptop computer, a netbook, a set-top box (STB), a PDA, an entertainment media system, a cellular phone, a smart phone, a mobile device, a wearable device (e.g., a smart watch), a smart home device (e.g., a smart appliance), other smart devices, a web device, a network router, a network switch, a network bridge, or any machine capable of executing, sequentially or otherwise, the instructions 1408 specifying actions to be taken by the computer 1400. In addition, while only a single computer 1400 is shown, the term "machine" should also be taken to include a collection of machines that individually or jointly execute the instructions 1408 to perform any one or more of the methodologies discussed herein.

[0092] Computer 1400 may include processor 1402, memory 1404, and I / O components 1442 that may be configured to communicate with each other via 1444. In an example embodiment, processor 1402 (e.g., a central processing unit (CPU), a reduced instruction set computing (RISC) processor, a complex instruction set computing (CISC) processor, a graphics processing unit (GPU), a digital signal processor (DSP), an ASIC, a radio frequency integrated circuit (RFIC), other processors, or any suitable combination thereof) may include, for example, processor 1406 that executes instructions 1408 and processor 1410. The term "processor" is intended to include multi-core processors that may include two or more independent processors (sometimes referred to as "cores") that may execute instructions simultaneously. Although Fig.14 Multiple processors 1402 are shown, but computer 1400 may include a single processor with a single core, a single processor with multiple cores (e.g., a multi-core processor), multiple processors with a single core, multiple processors with multiple cores, or any combination thereof.

[0093] The memory 1404 includes a main memory 1412, a static memory 1414, and a storage unit 1416, which are accessible by the processor 1402 via 1444. The main memory 1404, the static memory 1414, and the storage unit 1416 store instructions 1408 that embody any one or more of the methodologies or functions described herein. The instructions 1408 may also reside, in whole or in part, within the main memory 1412, within the static memory 1414, within the machine-readable medium 1418 within the storage unit 1416, within at least one of the processors 1402 (e.g., within a cache memory of the processor), or within any suitable combination thereof during execution of the instructions 1408 by the computer 1400.

[0094] I / O components 1442 may include a variety of components for receiving input, providing output, generating output, transmitting information, exchanging information, capturing measurements, etc. The specific I / O components 1442 included in a particular machine will depend on the type of machine. For example, a portable machine such as a mobile phone may include a touch input device or other such input mechanism, while a headless server machine will be less likely to include such a touch input device. It will be appreciated that I / O components 1442 may be included in Fig.14 Many other components not shown in the figure. In various example embodiments, I / O component 1442 may include 1428 and 1430. 1428 may include a visual component (e.g., a display such as a plasma display panel (PDP), a light emitting diode (LED) display, a liquid crystal display (LCD), a projector, or a cathode ray tube (CRT)), an acoustic component (e.g., a speaker), a tactile component (e.g., a vibration motor, a resistance mechanism), other signal generators, etc. 1430 may include an alphanumeric input component (e.g., a keyboard, a touch screen configured to receive alphanumeric input, an optical keyboard, or other alphanumeric input component), a point-based input component (e.g., a mouse, a touch pad, a trackball, a joystick, a motion sensor, or other pointing instrument), a tactile input component (e.g., a physical button, a touch screen or other tactile input component that provides the location and / or force of a touch or touch gesture), an audio input component (e.g., a microphone), etc.

[0095] In other example embodiments, I / O component 1442 may include 1432, 1434, 1436, or 1438 and various other components. For example, 1432 includes components for detecting expressions (e.g., hand expressions, facial expressions, voice expressions, body postures, or eye tracking), measuring bio-signals (e.g., blood pressure, heart rate, body temperature, sweating, or brain waves), identifying people (e.g., voice recognition, retinal recognition, facial recognition, fingerprint recognition, or EEG-based recognition), etc. 1434 includes an acceleration sensor component (e.g., an accelerometer), a gravity sensor component, a rotation sensor component (e.g., a gyroscope), etc. 1436 includes, for example, a light sensor component (e.g., a photometer), a temperature sensor component (e.g., one or more thermometers that detect ambient temperature), a humidity sensor component, a pressure sensor component (e.g., a barometer), an acoustic sensor component (e.g., one or more microphones that detect background noise), a proximity sensor component (e.g., an infrared sensor that detects nearby objects), a gas sensor (e.g., a gas detection sensor for detecting dangerous gas concentrations to ensure safety or measuring pollutants in the atmosphere), or other components that can provide indications, measurements, or signals corresponding to the surrounding physical environment. 1438 includes a position sensor component (e.g., a GPS receiver component), an altitude sensor component (e.g., an altimeter or barometer that detects air pressure from which altitude can be derived), an orientation sensor component (e.g., a magnetometer), etc.

[0096] A variety of technologies may be used to achieve communication. I / O components 1442 also include 1440, which is operable to couple computer 1400 to network 1420 or device 1422 via 1424 and 1426, respectively. For example, 1440 may include a network interface component or another suitable device to interface with network 1420. In other examples, 1440 may include a wired communication component, a wireless communication component, a cellular communication component, a near field communication (NFC) component, Parts (e.g. Low energy consumption), Components and other communication components for providing communication via other modalities. Device 1422 can be another machine or any peripheral device among various peripheral devices (e.g., a peripheral device coupled via USB).

[0097] In addition, 1440 can detect an identifier or include components operable to detect an identifier. For example, 1440 can include a radio frequency identification (RFID) tag reader component, an NFC smart tag detection component, an optical reader component (e.g., an optical sensor for detecting one-dimensional barcodes such as Universal Product Code (UPC) barcodes, multi-dimensional barcodes such as Quick Response (QR) codes, Aztec codes, Data Matrix, Dataglyph, MaxiCode, PDF417, UltraCode, UCC RSS-2D barcodes, and other optical codes), or an acoustic detection component (e.g., a microphone for identifying an audio signal of a tag). In addition, various information can be derived via 1440, such as a location via Internet Protocol (IP) geolocation, a location via Internet Protocol (IP), ... The location of signal triangulation, the location of an NFC beacon signal via detection that can indicate a specific location, etc.

[0098] Various memories (e.g., memory 1404, main memory 1412, static memory 1414, and / or memory of processor 1402) and / or storage unit 1416 may store one or more sets of instructions and data structures (e.g., software) that embody or are used by any one or more of the methods or functions described herein. These instructions (e.g., instructions 1408) when executed by processor 1402 cause various operations to implement the disclosed embodiments.

[0099] Instructions 1408 may be sent or received via a network interface device (e.g., a network interface component included in 1440) using a transmission medium and using any of a number of well-known transmission protocols (e.g., Hypertext Transfer Protocol (HTTP)) over network 1420. Similarly, instructions 1408 may be sent or received to device 1422 via 1426 (e.g., a peer-to-peer coupling) using a transmission medium.

[0100] Although the embodiments have been described with reference to specific example embodiments, it will be apparent that various modifications and changes may be made to these embodiments without departing from the broader scope of the present disclosure. Therefore, the description and the drawings should be considered in an illustrative sense, rather than a restrictive sense. The drawings forming part of the present invention show, by way of illustration and not limitation, specific embodiments in which the subject matter may be practiced. The embodiments shown are described in sufficient detail to enable those skilled in the art to practice the teachings disclosed herein. Other embodiments may be utilized and derived therefrom so that structural and logical substitutions and changes may be made without departing from the scope of the present disclosure. Therefore, the specific embodiments should not be understood in a restrictive sense, and the scope of the various embodiments is limited only by the appended claims and the full scope of the equivalents given by these claims.

[0101] These embodiments of the subject matter of the present invention may be referred to herein individually and / or collectively by the term "invention", which is merely for convenience and is not intended to voluntarily limit the scope of the present application to any single invention or inventive concept if, in fact, more than one invention or inventive concept is disclosed. Therefore, although specific embodiments have been shown and described herein, it should be understood that any arrangement calculated to achieve the same purpose may replace the specific embodiments shown. The present disclosure is intended to cover any and all adaptations or variations of the various embodiments. Upon reviewing the above description, the combination of the above embodiments and other embodiments not specifically described herein will be apparent to those skilled in the art.

[0102] An abstract of the disclosure is provided to allow the reader to quickly determine the nature of the technical disclosure. The abstract is submitted with the understanding that the abstract will not be used to interpret or limit the scope or meaning of the claims. In addition, in the foregoing detailed description, it can be seen that various features are combined together in a single embodiment for the purpose of simplifying the disclosure. This method of the disclosure should not be interpreted as reflecting the intention that the claimed embodiment requires more features than the features explicitly recited in each claim. On the contrary, as reflected in the attached claims, the subject matter of the present invention lies in less than all the features of a single disclosed embodiment. Therefore, the claims are incorporated into the detailed description here, with each claim independently serving as a separate embodiment.

[0103] Example

[0104] Example 1 is a method comprising: generating an image using a camera of a mobile device; acquiring a virtual object corresponding to an object in the image; identifying shading parameters of the virtual object based on the object captured in the image and a machine learning model pre-trained with a paired dataset, the paired dataset comprising synthetic source data and synthetic target data, the synthetic source data comprising an environment map and a three-dimensional (3D) scan of an object depicted in the environment map, the synthetic target data comprising a synthetic sphere image rendered in the same environment map; applying the shading parameters to the virtual object; and displaying the shaded virtual object as a layer of the image in a display of the mobile device.

[0105] Example 2 includes the method of Example 1, further comprising: obtaining synthetic source data; using a neural network to generate predicted lighting parameters based on the synthetic source data; using a differential renderer to generate a predicted spherical image based on the predicted lighting parameters; and using an L2 loss function to compare the predicted spherical image with the synthetic spherical image.

[0106] Example 3 includes the method of Example 1, wherein the environment map comprises a set of HDR environment atlases, wherein the 3D scans of the objects comprise a set of 3D facial scans of people depicted in the corresponding HDR environment maps.

[0107] Example 4 includes the method of Example 3, further comprising: training a machine learning model by: generating a synthetic facial image based on an HDR environment atlas and a set of 3D facial scans of a person using a first renderer; generating a synthetic sphere image based on the HDR environment atlas and a sphere asset using a second renderer; generating a predicted sphere image based on the synthetic facial image; and comparing the predicted sphere image to the synthetic sphere image using an L2 loss function.

[0108] Example 5 includes the method of Example 4, further comprising: predicting a spherical Gaussian and ambient light based on the synthesized facial image using a neural network; and generating a predicted spherical image based on the sphere asset, the spherical Gaussian, and the ambient light using a differential renderer.

[0109] Example 6 includes the method of Example 1, wherein applying the shading parameters to the virtual object comprises: providing the shading parameters to a physically based rendering (PBR) shader; and applying the estimated lighting conditions to the virtual object using the PBR shader.

[0110] Example 7 includes the method of Example 1, wherein the image comprises a self-portrait image of a user of the mobile device, and wherein the 3D scan of the item comprises a set of 3D facial scans of people depicted in the corresponding environment map.

[0111] Example 8 is a computing device comprising: a processor; and a memory storing instructions that, when executed by the processor, configure the computing device to: generate an image using a camera of a mobile device; acquire a virtual object corresponding to an object in the image; identify shading parameters of the virtual object based on the object captured in the image and a machine learning model pre-trained with a paired data set, the paired data set comprising synthetic source data and synthetic target data, the synthetic source data comprising an environment map and a three-dimensional (3D) scan of an object depicted in the environment map, the synthetic target data comprising a synthetic sphere image rendered in the same environment map; apply the shading parameters to the virtual object; and display the shaded virtual object as a layer of an image in a display of the mobile device.

[0112] Example 9 includes the computing device of Example 8, wherein the instructions further configure the computing device to: obtain synthetic source data; generate predicted lighting parameters based on the synthetic source data using a neural network; generate a predicted sphere image based on the predicted lighting parameters using a differential renderer; and compare the predicted sphere image to the synthetic sphere image using an L2 loss function.

[0113] Example 10 includes the computing device of Example 8, wherein the environment map comprises a set of HDR environment atlases, and wherein the 3D scans of the objects comprise a set of 3D facial scans of people depicted in the corresponding HDR environment maps.

[0114] Example 11 includes the computing device of Example 10, wherein the instructions further configure the computing device to: train the machine learning model by: generating a synthetic facial image based on an HDR environment atlas and a set of 3D facial scans of a person using a first renderer; generating a synthetic sphere image based on the HDR environment atlas and a sphere asset using a second renderer; generating a predicted sphere image based on the synthetic facial image; and comparing the predicted sphere image to the synthetic sphere image using an L2 loss function.

[0115] Example 12 includes the computing device of Example 11, wherein the instructions further configure the computing device to: predict a spherical Gaussian and ambient light based on the synthesized facial image using a neural network; and generate a predicted sphere image based on the sphere asset, the spherical Gaussian, and the ambient light using a differential renderer.

[0116] Example 13 includes the computing device of Example 8, wherein applying the shading parameters to the virtual object comprises: providing the shading parameters to a physically based rendering (PBR) shader; and applying the estimated lighting conditions to the virtual object using the PBR shader.

[0117] Example 14 includes the computing device of Example 8, wherein the image comprises a self-portrait image of a user of the mobile device, and wherein the 3D scan of the object comprises a set of 3D facial scans of people depicted in the corresponding environment map.

[0118] Example 15 is a non-transitory computer-readable storage medium, which includes instructions that, when executed by a computer, cause the computer to: generate an image using a camera of a mobile device; obtain a virtual object corresponding to an object in the image; identify shading parameters of the virtual object based on the object captured in the image and a machine learning model pre-trained with a paired dataset, the paired dataset including synthetic source data and synthetic target data, the synthetic source data including an environment map and a three-dimensional (3D) scan of an object depicted in the environment map, the synthetic target data including a synthetic sphere image rendered in the same environment map; apply the shading parameters to the virtual object; and display the shaded virtual object as a layer of the image in a display of the mobile device.

[0119] Example 16 includes the computer-readable storage medium of Example 15, wherein the instructions further cause the computer to: obtain synthetic source data; generate predicted lighting parameters based on the synthetic source data using a neural network; generate a predicted sphere image based on the predicted lighting parameters using a differential renderer; and compare the predicted sphere image to the synthetic sphere image using an L2 loss function.

[0120] Example 17 includes the computer-readable storage medium of Example 15, wherein the environment map comprises a set of HDR environment atlases, and wherein the 3D scans of the objects comprise a set of 3D facial scans of people depicted in the corresponding HDR environment maps.

[0121] Example 18 includes the computer-readable storage medium of Example 17, wherein the instructions further cause the computer to: train the machine learning model by: generating a synthetic facial image based on an HDR environment atlas and a set of 3D facial scans of a person using a first renderer; generating a synthetic sphere image based on the HDR environment atlas and a sphere asset using a second renderer; generating a predicted sphere image based on the synthetic facial image; and comparing the predicted sphere image to the synthetic sphere image using an L2 loss function.

[0122] Example 19 includes the computer-readable storage medium of Example 18, wherein the instructions further cause the computer to: predict a spherical Gaussian and ambient light based on the synthesized facial image using a neural network; and generate a predicted sphere image based on the sphere asset, the spherical Gaussian, and the ambient light using a differential renderer.

[0123] Example 20 includes the computer-readable storage medium of Example 15, wherein applying the shading parameters to the virtual object comprises: providing the shading parameters to a physically based rendering (PBR) shader; and applying the estimated lighting conditions to the virtual object using the PBR shader, wherein the image comprises a self-portrait image of a user of the mobile device, and wherein the 3D scan of the object comprises a set of 3D facial scans of people depicted in a corresponding environment map.

[0124] Glossary

[0125] "Carrier signal" refers to any intangible medium that can store, encode or carry instructions for execution by a machine and includes digital or analog communications signals or other intangible media to facilitate communication of such instructions. Instructions may be sent or received over a network using a transmission medium via a network interface device.

[0126] "Client Device" refers to any machine that interfaces with a communications network to obtain resources from one or more server systems or other client devices. A client device may be, but is not limited to, a mobile phone, a desktop computer, a laptop computer, a portable digital assistant (PDA), a smart phone, a tablet computer, an ultrabook, a netbook, multiple laptop computers, a multiprocessor system, a microprocessor-based or programmable consumer electronics product, a game console, a set-top box, or any other communications device that a user may use to access a network.

[0127] "Communications network" means one or more parts of a network, which may be an ad hoc network, an intranet, an extranet, a virtual private network (VPN), a local area network (LAN), a wireless LAN (WLAN), a wide area network (WAN), a wireless WAN (WWAN), a metropolitan area network (MAN), the Internet, a part of the Internet, a part of the public switched telephone network (PSTN), a plain old telephone service (POTS) network, a cellular telephone network, a wireless network, The coupling may be a network, another type of network, or a combination of two or more such networks. For example, the network or a portion of the network may include a wireless network or a cellular network, and the coupling may be a code division multiple access (CDMA) connection, a global system for mobile communications (GSM) connection, or other type of cellular or wireless coupling. In this example, the coupling may implement any of various types of data transmission technologies, such as single carrier radio transmission technology (1xRTT), evolution data optimized (EVDO) technology, general packet radio service (GPRS) technology, enhanced data rate for GSM evolution (EDGE) technology, including the third generation partnership project (3GPP) of 3G, fourth generation wireless (4G) network, universal mobile telecommunications system (UMTS), high speed packet access (HSPA), world wide interoperability for microwave access (WiMAX), long term evolution (LTE) standards, other data transmission technologies defined by various standard setting organizations, other long distance protocols, or other data transmission technologies.

[0128] "Component" refers to a device, physical entity or logic with the following boundaries: the boundaries are defined by function or subroutine calls, branch points, APIs or other technologies that provide partitioning or modularization of specific processing or control functions. Components can be combined with other components via their interfaces to perform machine processing. Components can be packaged functional hardware units designed for use with other components and can be part of a program that generally performs specific functions in related functions. Components can constitute software components (e.g., codes embodied on machine-readable media) or hardware components. "Hardware components" are tangible units that can perform certain operations and can be configured or arranged in a certain physical manner. In various examples, one or more computer systems (e.g., independent computer systems, client computer systems, or server computer systems) or one or more hardware components (e.g., processors or processor groups) of a computer system can be configured by software (e.g., applications or application parts) to operate to perform certain operations as described herein. Hardware components can also be implemented mechanically, electronically, or in any suitable combination thereof. For example, a hardware component may include a dedicated circuit system or logic that is permanently configured to perform certain operations. The hardware component may be a dedicated processor, such as a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC). The hardware component may also include a programmable logic or circuit system that is temporarily configured to perform certain operations by software. For example, the hardware component may include software executed by a general-purpose processor or other programmable processor. Once configured by such software, the hardware component becomes a specific machine (or a specific component of a machine) that is uniquely customized to perform the configured function and is no longer a general-purpose processor. It will be recognized that it can be determined for cost and time considerations whether to mechanically implement the hardware component in a dedicated and permanently configured circuit system or in a temporarily configured (e.g., configured by software) circuit system to implement the hardware component. Therefore, the phrase "hardware component" (or "hardware-implemented component") should be understood to include a tangible entity, that is, an entity that is physically constructed, permanently configured (e.g., hardwired) or temporarily configured (e.g., programmed) to operate in some way or perform certain operations described herein. Considering an example in which a hardware component is temporarily configured (e.g., programmed), it is not necessary to configure or instantiate each hardware component in the hardware component at any one time. For example, where a hardware component includes a general purpose processor that is configured by software to be a special purpose processor, the general purpose processor may be configured at different times to be respectively different special purpose processors (e.g., including different hardware components). The software configures a specific one or more processors accordingly, for example, to constitute a specific hardware component at one time and to constitute different hardware components at different times. Hardware components may provide information to other hardware components and receive information from other hardware components. Therefore, the described hardware components may be considered to be communicatively coupled.In the case of multiple hardware components being present at the same time, communication can be realized by (for example, by appropriate circuits and buses) transmitting signals between or among two or more hardware components in these hardware components. In examples where multiple hardware components are configured or instantiated at different times, communication between such hardware components can be realized, for example, by storing information in a memory structure accessible to multiple hardware components and retrieving information in the memory structure. For example, a hardware component can perform an operation, and the output of the operation is stored in a memory device coupled to it in communication. Then, other hardware components can access the memory device at a subsequent time to retrieve the stored output and process it. The hardware component can also initiate communication with an input device or an output device, and can operate on resources (for example, a collection of information). The various operations of the example methods described in this article can be performed at least in part by one or more processors that are temporarily configured (for example, by software) or permanently configured to perform related operations. Whether it is temporarily configured or permanently configured, such a processor can constitute a processor-implemented component that operates to perform one or more operations or functions described herein. As used herein, "processor-implemented components" refer to hardware components implemented using one or more processors. Similarly, the method described in this article can be implemented at least in part by a processor, wherein one or more specific processors are examples of hardware. For example, at least some operations of the method can be performed by one or more processors or the parts implemented by the processor. In addition, one or more processors can also operate to support the execution of related operations in a "cloud computing" environment or as a "software as a service" (SaaS) operation. For example, at least some operations in the operation can be performed by a computer group (as an example of a machine including a processor), wherein these operations can be accessed via a network (for example, the Internet) and via one or more appropriate interfaces (for example, API). The execution of certain operations can be distributed between processors, not only resident in a single machine, but also deployed across multiple machines. In some examples, a processor or a part implemented by a processor can be located in a single geographic location (for example, in a home environment, an office environment, or a server cluster). In other examples, a processor or a part implemented by a processor can be distributed across multiple geographic locations.

[0129] "Computer-readable storage media" refers to both machine storage media and transmission media. Therefore, these terms include both storage devices / media and carrier / modulated data signals. The terms "machine-readable medium", "computer-readable medium" and "device-readable medium" mean the same thing and can be used interchangeably in this disclosure.

[0130] "Ephemeral messages" are messages that can be accessed for a limited duration of time. Ephemeral messages can be text, images, videos, etc. The access time of an ephemeral message can be set by the sender of the message. Alternatively, the access time can be a default setting or a setting specified by the recipient. Regardless of the setting technique, the message is ephemeral.

[0131] "Machine storage media" refers to a single or multiple storage devices and media (e.g., centralized or distributed databases, and associated caches and servers) that store executable instructions, routines, and data. Thus, the term should be considered to include, but is not limited to, solid-state memory and optical and magnetic media, including memory internal or external to the processor. Specific examples of machine storage media, computer storage media, and device storage media include: non-volatile memory, including, for example, semiconductor memory devices such as erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), FPGA, and flash memory devices; magnetic disks, such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The terms "machine storage media", "device storage media", "computer storage media" mean the same thing and may be used interchangeably in this disclosure. The terms "machine storage media", "computer storage media", and "device storage media" explicitly exclude carrier waves, modulated data signals, and other such media, at least some of which are covered by the term "signal media".

[0132] “Non-transitory computer-readable storage medium” refers to tangible media capable of storing, encoding, or carrying instructions to be executed by a machine.

[0133] "Signal medium" refers to any intangible medium that can store, encode or carry instructions executed by a machine and includes digital or analog communication signals or other intangible media that facilitate the communication of software or data. The term "signal medium" should be deemed to include any form of modulated data signals, carrier waves, etc. The term "modulated data signal" means a signal in which one or more of its characteristics is set or changed in such a manner as to encode information in the signal. The terms "transmission medium" and "signal medium" mean the same thing and may be used interchangeably in this disclosure.

Claims

1. A method comprising: generating an image using a camera device of a mobile device; Acquire a virtual object corresponding to the object in the image; identifying shading parameters of the virtual object based on the object captured in the image and a machine learning model pre-trained with a paired dataset, the paired dataset comprising synthetic source data and synthetic target data, the synthetic source data comprising an environment map and a three-dimensional (3D) scan of an object depicted in the environment map, the synthetic target data comprising a synthetic sphere image rendered in the same environment map; applying the shading parameters to the virtual object; as well as The rendered virtual object is displayed as a layer of the image in a display of the mobile device.

2. The method according to claim 1, further comprising: Acquire the synthetic source data; generating predicted lighting parameters based on the synthetic source data using a neural network; generating a predicted spherical image based on the predicted lighting parameters using a differential renderer; as well as The predicted sphere image is compared with the synthesized sphere image using an L2 loss function.

3. The method according to claim 1, wherein: The environment map comprises a set of HDR environment atlases, wherein the 3D scans of the objects comprise a set of 3D facial scans of persons depicted in the corresponding HDR environment maps.

4. The method according to claim 3, further comprising: Train the machine learning model by: generating a synthetic facial image based on the HDR environment atlas and the set of 3D facial scans of the person using a first renderer; generating a composite sphere image based on the HDR environment atlas and the sphere asset using a second renderer; generating a predicted spherical image based on the synthesized facial image; as well as The predicted sphere image is compared with the synthesized sphere image using an L2 loss function.

5. The method according to claim 4, further comprising: predicting a spherical Gaussian and ambient light based on the synthesized facial image using a neural network; as well as The predicted sphere image is generated using a differential renderer based on the sphere asset, the spherical Gaussian, and ambient light.

6. The method according to claim 1, wherein: Applying the shading parameters to the virtual object comprises: providing the shading parameters to a physically based rendering (PBR) shader; and The estimated lighting conditions are applied to the virtual object using the PBR shader.

7. The method according to claim 1, wherein: The image comprises a self-portrait image of a user of the mobile device, wherein the 3D scan of the item comprises a set of 3D facial scans of people depicted in a corresponding environment map.

8. A computing device, comprising: processor; as well as a memory storing instructions that, when executed by the processor, configure the computing device to: generating an image using a camera device of a mobile device; Acquire a virtual object corresponding to the object in the image; identifying shading parameters of the virtual object based on the object captured in the image and a machine learning model pre-trained with a paired dataset, the paired dataset comprising synthetic source data and synthetic target data, the synthetic source data comprising an environment map and a three-dimensional (3D) scan of an object depicted in the environment map, the synthetic target data comprising a synthetic sphere image rendered in the same environment map; applying the shading parameters to the virtual object; as well as The rendered virtual object is displayed as a layer of the image in a display of the mobile device.

9. The computing device according to claim 8, wherein: The instructions further configure the computing device to: Acquire the synthetic source data; generating predicted lighting parameters based on the synthetic source data using a neural network; generating a predicted spherical image based on the predicted lighting parameters using a differential renderer; as well as The predicted sphere image is compared with the synthesized sphere image using an L2 loss function.

10. The computing device according to claim 8, wherein: The environment map comprises a set of HDR environment atlases, wherein the 3D scans of the objects comprise a set of 3D facial scans of people depicted in the corresponding HDR environment maps.

11. The computing device according to claim 10, wherein: The instructions further configure the computing device to: Train the machine learning model by: generating a synthetic facial image based on the HDR environment atlas and the set of 3D facial scans of the person using a first renderer; generating a composite sphere image based on the HDR environment atlas and the sphere asset using a second renderer; generating a predicted spherical image based on the synthesized facial image; as well as The predicted sphere image is compared with the synthesized sphere image using an L2 loss function.

12. The computing device according to claim 11, wherein: The instructions further configure the computing device to: predicting spherical Gaussian and ambient light based on the synthesized facial image using a neural network; and The predicted sphere image is generated using a differential renderer based on the sphere asset, the spherical Gaussian, and ambient light.

13. The computing device of claim 8, wherein: Applying the shading parameters to the virtual object comprises: providing the shading parameters to a physically based rendering (PBR) shader; and The estimated lighting conditions are applied to the virtual object using the PBR shader.

14. The computing device of claim 8, wherein: The image comprises a self-portrait image of a user of the mobile device, wherein the 3D scan of the item comprises a set of 3D facial scans of people depicted in a corresponding environment map.

15. A non-transitory computer-readable storage medium comprising instructions that, when executed by a computer, cause the computer to: generating an image using a camera device of a mobile device; Acquire a virtual object corresponding to the object in the image; identifying shading parameters of the virtual object based on the object captured in the image and a machine learning model pre-trained with a paired dataset, the paired dataset comprising synthetic source data and synthetic target data, the synthetic source data comprising an environment map and a three-dimensional (3D) scan of an object depicted in the environment map, the synthetic target data comprising a synthetic sphere image rendered in the same environment map; applying the shading parameters to the virtual object; as well as The rendered virtual object is displayed as a layer of the image in a display of the mobile device.

16. The computer-readable storage medium of claim 15, wherein: The instructions further cause the computer to: Acquire the synthetic source data; generating predicted lighting parameters based on the synthetic source data using a neural network; generating a predicted spherical image based on the predicted lighting parameters using a differential renderer; as well as The predicted sphere image is compared with the synthesized sphere image using an L2 loss function.

17. The computer-readable storage medium of claim 15, wherein: The environment map comprises a set of HDR environment atlases, wherein the 3D scans of the objects comprise a set of 3D facial scans of people depicted in the corresponding HDR environment maps.

18. The computer-readable storage medium of claim 17, wherein: The instructions further cause the computer to: Train the machine learning model by: generating a synthetic facial image based on the HDR environment atlas and the set of 3D facial scans of the person using a first renderer; generating a composite sphere image based on the HDR environment atlas and the sphere asset using a second renderer; generating a predicted spherical image based on the synthesized facial image; as well as The predicted sphere image is compared with the synthesized sphere image using an L2 loss function.

19. The computer-readable storage medium of claim 18, wherein: The instructions further cause the computer to: predicting spherical Gaussian and ambient light based on the synthesized facial image using a neural network; and The predicted sphere image is generated using a differential renderer based on the sphere asset, the spherical Gaussian, and ambient light.

20. The computer-readable storage medium of claim 15, wherein: Applying the shading parameters to the virtual object comprises: providing the shading parameters to a physically based rendering (PBR) shader; and applying estimated lighting conditions to the virtual object using the PBR shader, Wherein the image comprises a self-portrait image of a user of the mobile device, and wherein the 3D scan of the item comprises a set of 3D facial scans of people depicted in a corresponding environment map.

Images