A method and electronic device for image signal processing by using compressed domain artificial intelligence

Abstract

A method and an electronic device for image signal processing are provided. The method includes obtaining a raw image and mapping, using a raw image encoder, the raw image to a compressed domain. The raw image is represented using latent variables in the compressed domain. The method includes performing one or more image signal processing operations on the latent variables. The method includes mapping, using an output image decoder, the processed latent variables to an output image in an output color space.

Description

A METHOD AND ELECTRONIC DEVICE FOR IMAGE SIGNAL PROCESSING BY USING COMPRESSED DOMAIN ARTIFICIAL INTELLIGENCE

[0001] This disclosure relates to image signal processing. More specifically, this disclosure relates to a method and an electronic device for image signal processing by using compressed domain artificial intelligence.

[0002] As imaging resolutions of electronic devices, such as mobile devices, have increased, memory required to process these images has also significantly grown. This has led to a larger amount of data to be handled for high-resolution images, but the available memory in such devices is insufficient, thereby resulting in slower processing speeds or degraded performance. Consequently, memory requirements have become a bottleneck in the development of more performant electronic devices.

[0003] One way to address that problem is to compress high-resolution images to a compressed domain with smaller size and perform downstream tasks, such as demosaicing and denoising, in this domain. However, designing an effective and general compression domain for downstream tasks remains challenging, and current techniques in image compression have limitations. For instance, most image compression processes are designed for red-green-blue (RGB) images rather than raw images, while raw data processing is essential in modern image signal processing (ISP). Applying image compression to raw data tends to reach a local minimum solution with inferior performance. Also, existing image compression techniques are not directly designed for compressed domain ISP downstream tasks.

[0004] According to an embodiment of the disclosure, provided is a method for image signal processing. The method may include obtaining a raw image. The method may include mapping, using a raw image encoder, the raw image to a compressed domain. The raw image may be represented using latent variables in the compressed domain. The method may include performing one or more image signal processing operations on the latent variables, where (i) each of the one or more image signal processing operations is configured to operate in the compressed domain and (ii) the one or more image signal processing operations generate processed latent variables. The method may include mapping, using an output image decoder, the processed latent variables to an output image in an output color space.

[0005] According to an embodiment of the disclosure, provided is an electronic device for image signal processing. The electronic device may include at least one processing device. The at least one processing device may be configured to obtain a raw image. The at least one processing device may be configured to map, using a raw image encoder, the raw image to a compressed domain. The raw image is represented using latent variables in the compressed domain. The at least one processing device may be configured to perform one or more image signal processing operations on the latent variables, where (i) each of the one or more image signal processing operations is configured to operate in the compressed domain and (ii) the one or more image signal processing operations are configured to generate processed latent variables. The at least one processing device may be configured to map, using an output image decoder, the processed latent variables to an output image in an output color space.

[0006] According to an embodiment of the disclosure, a computer-readable storage medium storing instructions is provided. The instructions, when executed by at least one processor, may cause the at least one processor to perform the method for image signal processing.

[0007] Other technical features may be readily apparent to one skilled in the art from the following figures, descriptions, and claims.

[0008] For a more complete understanding of this disclosure and its advantages, reference is now made to the following description taken in conjunction with the accompanying drawings, in which like reference numerals represent like parts:

[0009] FIGURE 1 illustrates an example of a network configuration providing compressed domain artificial intelligence (AI) for image signal processing (ISP) in accordance with an embodiment of the disclosure;

[0010] FIGURE 2 illustrates a method for image signal processing by using a compressed domain AI in accordance with an embodiment of the disclosure;

[0011] FIGURE 3 illustrates an example of an architecture for derivation of a compressed domain for an AI model in accordance with an embodiment of the disclosure;

[0012] FIGURE 4 illustrates an example of an architecture for a compressed domain for a number of different data formats for an AI model in accordance with an embodiment of the disclosure;

[0013] FIGURE 5 illustrates an example of an architecture for operation of an AI model within a compressed domain in accordance with an embodiment of the disclosure;

[0014] FIGURE 6 illustrates an example of an AI model for efficient raw domain compressed image data ISP in accordance with an embodiment of the disclosure;

[0015] FIGURE 7 illustrates an example of an architecture for progressive compression-decompression for an AI model operating in the compressed domain in accordance with an embodiment of the disclosure;

[0016] FIGURES 8A through 8C comparatively illustrate example compression and inherent demosaicing for raw data when using a compressed domain AI for image signal processing in accordance with an embodiment of the disclosure;

[0017] FIGURES 9A through 9C comparatively illustrate example compression and inherent demosaicing for raw data when using a compressed domain AI for image signal processing using an enlargement of a portion of the image content from FIGURES 8A through 8C in accordance with an embodiment of the disclosure;

[0018] FIGURES 10A through 10C comparatively illustrate an example of the result of raw image denoising in the compressed latent domain for image signal processing in accordance with an embodiment of the disclosure;

[0019] FIGURES 11A through 11C comparatively illustrate an example of the result of multi-frame blending in the compressed domain for image signal processing in accordance with an embodiment of the disclosure;

[0020] FIGURES 12A and 12B illustrate an example of the result of progressive encoding and decoding in accordance with an embodiment of the disclosure;

[0021] FIGURE 13 illustrates an example of AI model for efficient raw domain compressed image data ISP in accordance with an embodiment of the disclosure;

[0022] FIGURE 14 illustrates an example of an architecture for generalized progressive compression-decompression for an AI model operating in the compressed domain in accordance with an embodiment of the disclosure;

[0023] FIGURE 15 illustrates an example of an architecture for derivation of a compressed domain for an AI model utilizing generalized progressive compression-decompression in accordance with an embodiment of the disclosure;

[0024] FIGURE 16 illustrates an example of an architecture for generalized progressive compression-decompression for an AI model in accordance with an embodiment of the disclosure;

[0025] FIGURES 17A through 17C comparatively illustrate example compression and inherent demosaicing for raw data when using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with an embodiment of the disclosure;

[0026] FIGURES 18A through 18C comparatively illustrate example raw image denoising using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with an embodiment of the disclosure;

[0027] FIGURES 19A through 19C comparatively illustrate example multi-frame blending using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with an embodiment of the disclosure;

[0028] FIGURES 20A through 20C comparatively illustrate example image registration using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with an embodiment of the disclosure;

[0029] FIGURE 21 illustrates a multi-frame ISP pipeline;

[0030] FIGURE 22 illustrates an example of a multi-frame ISP pipeline;

[0031] FIGURE 23 illustrates a pipeline for the ISP sequence of FIGURE 22 in the context of a compute engine and memory;

[0032] FIGURE 24 illustrates a pipeline for the ISP sequence of FIGURE 22 in the context of a compute engine and memory and assuming a specified data compression ratio;

[0033] FIGURE 25 illustrates an example of a pipeline utilizing AI to compress data and performing AI-based processing on the compressed data in accordance with an embodiment of the disclosure;

[0034] FIGURE 26 illustrates an example of a pipeline utilizing AI to compress data and performing AI-based processing on the compressed data in accordance with an embodiment of the disclosure;

[0035] FIGURE 27 illustrates an example of a compressed domain AI multi-frame multi-task pipeline in accordance with an embodiment of the disclosure;

[0036] FIGURE 28 illustrates an example of a pipeline utilizing progressive AI compression and using most significant bits to efficiently perform downstream tasks that do not need full-resolution data in accordance with an embodiment of the disclosure;

[0037] FIGURE 29 illustrates an example of a pipeline utilizing progressive AI compression and using most significant bits to efficiently perform downstream tasks that do not need full-resolution data in accordance with an embodiment of the disclosure;

[0038] FIGURE 30 illustrates an example of a pipeline utilizing AI to compress multi-frame data using temporal redundancy between frames in accordance with an embodiment of the disclosure;

[0039] FIGURES 31A through 31C comparatively illustrate an example of a result of multi-frame blending in a compressed domain for AI-based image signal processing in accordance with an embodiment of the disclosure;

[0040] FIGURES 32A through 32D comparatively illustrate an example of a result of compressed domain for AI denoising in accordance with an embodiment of the disclosure; and

[0041] FIGURES 33A through 33C comparatively illustrate an example of a result of compressed domain for AI registration in accordance with an embodiment of the disclosure.

[0042] The terms used in the present disclosure will be briefly described, and then embodiments of the present disclosure will be described in detail.

[0043] The terms used in the present disclosure are general terms as much as possible and have been widely used nowadays in consideration of the functions of the present disclosure, which, however, may be changed according to an intention of a technician in the art, a precedent, the advent of new technologies, or the like. Also, particular cases may include terms arbitrary selected by an applicant, and in this case, the meaning of the terms will be described in detail in the corresponding description. Therefore, the terms used in the present disclosure should be defined based on the meanings of the terms and the content throughout the present disclosure, rather than simply based on the titles of the terms.

[0044] It may be advantageous to set forth definitions of certain words and phrases used throughout this disclosure. Thus, he terms "transmit", "receive", and "communicate," as well as derivatives thereof, encompass both direct and indirect communication. The terms "include" and "comprise", as well as derivatives thereof, mean inclusion without limitation. The term "or" is inclusive, meaning and / or. The phrase "associated with", as well as derivatives thereof, means to include, be included within, interconnect with, contain, be contained within, connect to or with, couple to or with, be communicable with, cooperate with, interleave, juxtapose, be proximate to, be bound to or with, have, have a property of, have a relationship to or with, or the like.

[0045] Moreover, various functions described below may be implemented or supported by one or more computer programs, each of which is formed from a computer readable program code and embodied in a computer readable medium. The terms "application" and "program" refer to one or more computer programs, software components, sets of instructions, procedures, functions, objects, classes, instances, related data, or a portion thereof adapted for implementation in a suitable computer readable program code. The phrase "computer readable program code" includes any type of computer code, including source code, object code, and executable code. The phrase "computer readable medium"includes any type of medium capable of being accessed by a computer, such as read only memory (ROM), random access memory (RAM), a hard disk drive, a compact disc (CD), a digital video disc (DVD), or any other type of memory. A "non-transitory"computer readable medium excludes wired, wireless, optical, or other communication links that transport transitory electrical or other signals. A non-transitory computer readable medium includes media where data can be permanently stored and media where data can be stored and later overwritten, such as a rewritable optical disc or an erasable memory device.

[0046] As used here, terms and phrases such as "have", "may have", "include", or "may include" a feature (like a number, function, operation, or component such as a part) indicate the existence of the feature and do not exclude the existence of other features. Also, as used here, the phrases "A or B", "at least one of A and / or B", or "one or more of A and / or B" may include all possible combinations of A and B. For example, "A or B", "at least one of A and B", and "at least one of A or B" may indicate all of (1) including at least one A, (2) including at least one B, or (3) including at least one A and at least one B. Further, as used here, the terms "first" and "second" may modify various components regardless of importance and do not limit the components. These terms are only used to distinguish one component from another. For example, a first user device and a second user device may indicate different user devices from each other, regardless of the order or importance of the devices. A first component may be denoted a second component and vice versa without departing from the scope of this disclosure.

[0047] It will be understood that, when an element (such as a first element) is referred to as being (operatively or communicatively) "coupled with / to" or "connected with / to" another element (such as a second element), the element can be coupled or connected with / to the other element directly or via a third element. In contrast, it will be understood that, when an element (such as a first element) is referred to as being "directly coupled with / to" or "directly connected with / to" another element (such as a second element), no other element (such as a third element) intervenes between the element and the other element.

[0048] As used here, the phrase "configured (or set) to" may be interchangeably used with the phrases "suitable for", "having the capacity to", "designed to", "adapted to", "made to",or "capable of" depending on the circumstances. The phrase "configured (or set) to" does not essentially mean "specifically designed in hardware to". Rather, the phrase "configured to" may mean that a device can perform an operation together with another device or parts. For example, the phrase "processor configured (or set) to perform A, B, and C" may mean a generic-purpose processor (such as a CPU or application processor) that may perform the operations by executing one or more software programs stored in a memory device or a dedicated processor (such as an embedded processor) for performing the operations.

[0049] The terms and phrases as used here are provided merely to describe some embodiments of this disclosure but not to limit the scope of other embodiments of this disclosure. It is to be understood that the singular forms "a", "an", and "the" include plural references unless the context clearly dictates otherwise. All terms and phrases, including technical and scientific terms and phrases, used here have the same meanings as commonly understood by one of ordinary skill in the art to which the embodiments of this disclosure belong. It will be further understood that terms and phrases, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined here. In some cases, the terms and phrases defined here may be interpreted to exclude embodiments of this disclosure.

[0050] Examples of an "electronic device"according to embodiments of this disclosure may include at least one of a smartphone, a tablet personal computer (PC), a mobile phone, a video phone, an e-book reader, a desktop PC, a laptop computer, a netbook computer, a workstation, a personal digital assistant (PDA), a portable multimedia player (PMP), an MP3 player, a mobile medical device, a camera, or a wearable device (such as smart glasses, a head-mounted device (HMD), electronic clothes, an electronic bracelet, an electronic necklace, an electronic accessory, an electronic tattoo, a smart mirror, or a smart watch). Other examples of the "electronic device" include a smart home appliance. Examples of the smart home appliance may include at least one of a television, a digital video disc (DVD) player, an audio player, a refrigerator, an air conditioner, a cleaner, an oven, a microwave oven, a washer, a dryer, an air cleaner, a set-top box, a home automation control panel, a security control panel, a TV box (such as SAMSUNG HOMESYNC, APPLETV, or GOOGLE TV), a smart speaker or speaker with an integrated digital assistant (such as SAMSUNG GALAXY HOME, APPLE HOMEPOD, or AMAZON ECHO), a gaming console (such as an XBOX, PLAYSTATION, or NINTENDO), an electronic dictionary, an electronic key, a camcorder, or an electronic picture frame. Still other examples of the "electronic device" include at least one of various medical devices (such as diverse portable medical measuring devices (like a blood sugar measuring device, a heartbeat measuring device, or a body temperature measuring device), a magnetic resource angiography (MRA) device, a magnetic resource imaging (MRI) device, a computed tomography (CT) device, an imaging device, or an ultrasonic device), a navigation device, a global positioning system (GPS) receiver, an event data recorder (EDR), a flight data recorder (FDR), an automotive infotainment device, a sailing electronic device (such as a sailing navigation device or a gyro compass), avionics, security devices, vehicular head units, industrial or home robots, automatic teller machines (ATMs), point of sales (POS) devices, or Internet of Things (IoT) devices (such as a bulb, various sensors, electric or gas meter, sprinkler, fire alarm, thermostat, street light, toaster, fitness equipment, hot water tank, heater, or boiler). Other examples of the "electronic device" include at least one part of a piece of furniture or building / structure, an electronic board, an electronic signature receiving device, a projector, or various measurement devices (such as devices for measuring water, electricity, gas, or electromagnetic waves). Note that, according to various embodiments of this disclosure, the "electronic device" may be one or a combination of the above-listed devices. According to some embodiments of this disclosure, the "electronic device" may be a flexible electronic device. The electronic device disclosed here is not limited to the above-listed devices and may include new electronic devices depending on the development of technology.

[0051] In the following description, electronic devices are described with reference to the accompanying drawings, according to various embodiments of this disclosure. As used here, the term "user" may denote a human or another device (such as an artificial intelligent electronic device) using the electronic device.

[0052] Definitions for other certain words and phrases may be provided throughout this patent document. Those of ordinary skill in the art should understand that in many if not most instances, such definitions apply to prior as well as future uses of such defined words and phrases.

[0053] Throughout the present disclosure, expressions such as "at least one of a, b or c", "at least one of a, b, or c", "at least one of a, b and c", "at least one of a, b, and c" may indicate "a," "b," "c," "a and b," "a and c," "b and c," "all of a, b, and c," or variations thereof. Similarly, expressions such as "at least one of a or b", "at least one of a, or b", "at least one of a and b", "at least one of a, and b " may indicate "a," "b,", "a and b," or variations thereof.

[0054] In this disclosure, the expression "and / or" includes a combination of a plurality of described components or any component of the plurality of described components. In this disclosure, terms such as "1st," "2nd," "first," and "second" may be merely used to distinguish a corresponding component from other corresponding components and do not limit the corresponding components in terms of other aspects (for example, the degree of importance or the order).

[0055] Throughout the present disclosure, when a part "includes" or "comprises" an element, the part may further include other elements, not excluding the other elements, unless there is a particular description contrary thereto.

[0056] Also, the terms "portion," "module," etc. described in this disclosure denote a unit configured to process at least one function or operation, and the "portion," and the "module" may be realized as hardware or software, such as a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC), or a combination of the hardware and the software. The term "portion" used in an embodiment of the present disclosure does not have a meaning limited to software or hardware. A "portion" described in the present disclosure may be configured to be in a storage medium which may be addressed or may be configured to play one or more processors. According to an embodiment of the present disclosure, a "portion" may include components, such as software components, object-oriented software components, class components, and task components, processes, functions, attributes, procedures, sub-routines, segments of a program code, drivers, firmware, a microcode, a circuit, data, a database, data structures, tables, arrays, and variables. Functions provided through a predetermined component or a predetermined "portion" may be combined to reduce the number of functions or may be divided into additional components. Also, according to an embodiment, a "portion" may include one or more processors.

[0057] According to an embodiment of the present disclosure, each of blocks of the flowcharts and combinations of the flowcharts may be performed by computer program instructions. The computer program instructions may be loaded on a general-purpose computer, a specialized computer, or a processor of other programmable data processing device. The instructions performed through the computer or the processor of the other programmable data processing device may generate a medium for performing the functions described in the flowchart block(s). The computer program instructions may also be stored in a computer-available or computer-readable memory oriented for the computer or the other programmable data processing device in order to realize the functions in a predetermined way. The instructions stored in the computer-available or computer-readable memory may also produce manufacturing items embedding an instruction medium for performing the functions described in the flowchart block(s). The computer program instructions may also be loaded on the computer or the other programmable data processing device.

[0058] Furthermore, each block in the flowcharts may represent a module, segment, or a part of code including one or more executable instructions to perform particular logic function(s). According to an embodiment of the present disclosure, the functions described with respect to the blocks may also be generated not according to an order. For example, two blocks illustrated in succession may be executed substantially concurrently, or the blocks may sometimes be executed in a reverse order, depending on the functions involved therein.

[0059] Hereinafter,embodiments of the present disclosure will be described in detail with reference to the accompanying drawings, so that the embodiments of the present disclosure may be easily implemented by one of ordinary skill in the art. However, the embodiments of the present disclosure may have different forms and should not be construed as being limited to the embodiments of the present disclosure described herein. Also, in the drawings, parts not related to descriptions are omitted for the clear description of the embodiments of the present disclosure, and throughout the specification, like reference numerals are used for like elements.

[0060] This disclosure relates to a method and an electronic device for image signal processing by using compressed domain artificial intelligence.

[0061] FIGURES 1 through 33C, discussed below, and the various embodiments of this disclosure are described with reference to the accompanying drawings. However, it should be appreciated that this disclosure is not limited to these embodiments, and all changes and / or equivalents or replacements thereto also belong to the scope of this disclosure. The same or similar reference denotations may be used to refer to the same or similar elements throughout the specification and the drawings.

[0062] FIGURE 1 illustrates an example of a network configuration 100 providing compressed domain AI for image signal processing in accordance with an embodiment of the disclosure. The embodiment of the network configuration 100 shown in FIGURE 1 is for illustration only. Other embodiments of the network configuration 100 may be used without departing from the scope of this disclosure.

[0063] According to an embodiment of the disclosure, an electronic device 101 may be included in the network configuration 100. The electronic device 101 may include at least one of a bus 110, a processor 120, a memory 130, an input / output (I / O) interface 150, a display 160, a communication interface 170, or a sensor 180. In an embodiment, the electronic device 101 may exclude at least one of these components or may add at least one other component. The bus 110 may include a circuit for connecting the components 120, 130, 140, 150, 160, 170, and 180 with one another and for transferring communications (such as control messages and / or data) between the components 120, 130, 140, 150, 160, 170, and 180.

[0064] The processor 120 may include at least one processing device, such as one or more microprocessors, microcontrollers, digital signal processors (DSPs), application specific integrated circuits (ASICs), or field programmable gate arrays (FPGAs). In an embodiment, the processor 120 may include at least one of a central processing unit (CPU), an application processor (AP), a communication processor (CP), or a graphics processor unit (GPU). The processor 120 may be able to perform control on at least one of the other components of the electronic device 101 and / or perform an operation or data processing relating to communication or other functions. As described in more detail below, the processor 120 may perform various operations related to compressed domain artificial intelligence (AI) for image signal processing (ISP).

[0065] The memory 130 may include volatile and / or non-volatile memory. For example, the memory 130 may store commands or data related to at least one other component of the electronic device 101. According to an embodiment of the disclosure, the memory 130 may store software and / or a program 140. The program 140 may include, for example, a kernel 141, middleware 143, an application programming interface (API) 145, and / or an application program (or "application") 147. At least a portion of the kernel 141, middleware 143, or API 145 may be denoted an operating system (OS).

[0066] The kernel 141 may control or manage system resources (such as the bus 110, processor 120, or memory 130) used to perform operations or functions implemented in other programs (such as the middleware 143, API 145, or application 147). The kernel 141 may provide an interface that allows the middleware 143, the API 145, or the application 147 to access the individual components of the electronic device 101 to control or manage the system resources. The application 147 may support various functions related to compressed domain AI for image signal processing. These functions may be performed by a single application or by multiple applications that each carries out one or more of these functions. The middleware 143 may function as a relay to allow the API 145 or the application 147 to communicate data with the kernel 141, for instance. A plurality of applications 147 may be provided. The middleware 143 may be able to control work requests received from the applications 147, such as by allocating the priority of using the system resources of the electronic device 101 (like the bus 110, the processor 120, or the memory 130) to at least one of the plurality of applications 147. The API 145 may be an interface allowing the application 147 to control functions provided from the kernel 141 or the middleware 143. For example, the API 145 may include at least one interface or function (such as a command) for filing control, window control, image processing, or text control.

[0067] The I / O interface 150 may serve as an interface that can, for example, transfer commands or data input from a user or other external devices to other component(s) of the electronic device 101. The I / O interface 150 may output commands or data received from other component(s) of the electronic device 101 to the user or the other external device.

[0068] The display 160 may include, for example, a liquid crystal display (LCD), a light emitting diode (LED) display, an organic light emitting diode (OLED) display, a quantum-dot light emitting diode (QLED) display, a microelectromechanical systems (MEMS) display, or an electronic paper display. The display 160 may be a depth-aware display, such as a multi-focal display. The display 160 may be able to display, for example, various contents (such as text, images, videos, icons, or symbols) to the user. The display 160 may include a touchscreen and may receive, for example, a touch, gesture, proximity, or a hovering input using an electronic pen or a body portion of the user.

[0069] The communication interface 170, for example, may be able to set up communication between the electronic device 101 and an external electronic device (such as a first electronic device 102, a second electronic device 104, or a server 106). For example, the communication interface 170 may be connected with a network 162 or 164 through wireless or wired communication to communicate with the external electronic device. The communication interface 170 may be a wired or wireless transceiver or any other component for transmitting and receiving signals.

[0070] The wireless communication may be able to use at least one of, for example, WiFi, long term evolution (LTE), long term evolution-advanced (LTE-A), 5th generation wireless system (5G), millimeter-wave or 60 GHz wireless communication, Wireless USB, code division multiple access (CDMA), wideband code division multiple access (WCDMA), universal mobile telecommunication system (UMTS), wireless broadband (WiBro), or global system for mobile communication (GSM), as a communication protocol. The wired connection may include, for example, at least one of a universal serial bus (USB), high-definition multimedia interface (HDMI), recommended standard 232 (RS-232), or plain old telephone service (POTS). The network 162 or 164 may include at least one communication network, such as a computer network (like a local area network (LAN) or wide area network (WAN)), Internet, or a telephone network.

[0071] The electronic device 101 may include at least one sensor 180 that may measure a physical quantity or detect an activation state of the electronic device 101 and convert measured or detected information into an electrical signal. For example, at least one sensor 180 may include at least one camera or other imaging sensors for capturing images of scenes. The sensor(s) 180 may include at least one of one or more buttons for touch input, one or more microphones, a gesture sensor, a gyroscope or gyro sensor, an air pressure sensor, a magnetic sensor or magnetometer, an acceleration sensor or accelerometer, a grip sensor, a proximity sensor, a color sensor (such as an RGB sensor), a bio-physical sensor, a temperature sensor, a humidity sensor, an illumination sensor, an ultraviolet (UV) sensor, an electromyography (EMG) sensor, an electroencephalogram (EEG) sensor, an electrocardiogram (ECG) sensor, an infrared (IR) sensor, an ultrasound sensor, an iris sensor, or a fingerprint sensor. The sensor(s) 180 can further include an inertial measurement unit, which may include one or more accelerometers, gyroscopes, and other components. The sensor(s) 180 may include a control circuit for controlling at least one of the sensors included here. Any of these sensor(s) 180 may be located within the electronic device 101.

[0072] In an embodiment, the first external electronic device 102 or the second external electronic device 104 may be a wearable device or an electronic device-mountable wearable device (such as a head mounted display (or "HMD"). When the electronic device 101 may be mounted in the first external electronic device 102 (such as the HMD), the electronic device 101 may communicate with the first external electronic device 102 through the communication interface 170. The electronic device 101 may be directly connected with the first external electronic device 102 to communicate with the first external electronic device 102 without involving with a separate network. The electronic device 101 may be an augmented reality wearable device, such as eyeglasses, which include one or more imaging sensors, or a VR or XR headset.

[0073] The first and second external electronic devices 102 and 104 and the server 106 each may be a device of the same or a different type from the electronic device 101. According to an embodiment of the disclosure, the server 106 may include a group of one or more servers. Also, according to an embodiment of the disclosure, all or some of the operations executed on the electronic device 101 may be executed on another or multiple other electronic devices (such as the first external electronic devices 102 and the second external electronic device 104 or server 106). According to an embodiment of the disclosure, when the electronic device 101 should perform some function or service automatically or at a request, the electronic device 101, instead of executing the function or service on its own or additionally, may request another device (such as the first external electronic device 102 and the second external electronic device 104 or server 106) to perform at least some functions associated therewith. The other electronic device (such as the first external electronic device 102 and the second external electronic device 104 or server 106) may be able to execute the requested functions or additional functions and transfer a result of the execution to the electronic device 101. The electronic device 101 may provide a requested function or service by processing the received result as it is or additionally. To that end, a cloud computing, distributed computing, or client-server computing technique may be used, for example. While FIGURE 1 shows that the electronic device 101 includes the communication interface 170 to communicate with the second external electronic device 104 or server 106 via the network 162, the electronic device 101 may be independently operated without a separate communication function according to an embodiment of the disclosure.

[0074] The server 106 may include the same or similar components 110, 120, 130, 140, 150, 160, 170, 180 as the electronic device 101 (or a suitable subset thereof). The server 106 may support the electronic device 101 by performing at least one of operations (or functions) implemented on the electronic device 101. For example, the server 106 may include a processing module or at least one processor (or at least one processor circuitry) that may support the processor 120 implemented in the electronic device 101. As described in more detail below, the server 106 may perform various operations related to compressed domain AI for image signal processing.

[0075] Although FIGURE 1 illustrates an example of a network configuration 100 including an electronic device 101, various changes may be made to FIGURE 1. For example, the network configuration 100 may include any number of each component in any suitable arrangement. Computing and communication systems come in a wide variety of configurations, and FIGURE 1 does not limit the scope of the disclosure to any particular configuration. While FIGURE 1 illustrates one operational environment in which various features disclosed in the disclosure can be used, these features could be used in any other suitable system.

[0076] FIGURE 2 illustrates a method 200 for image signal processing by using a compressed domain AI in accordance with an embodiment of the disclosure. For ease of explanation, the method 200 of FIGURE 2 is described as being performed using the electronic device 101 (or using the processor 120 of the electronic device 101) in the network configuration 100 of FIGURE 1. However, the method 200 may be performed using any other suitable device(s) and in any other suitable system(s). The method 200 may be referred to as a process.

[0077] As shown in FIGURE 2, the method 200 may start with obtaining a raw image (step 201). The image may have any suitable resolution and image format. In a case, the image may represent a Tetra format image, a (quad) Bayer format image, a hexa-deca Bayer ("hexa-deca") format image, or a red-green-blue-white (RGBW) format image. Using a raw image encoder, the raw image may be processed to a compressed domain in which the raw image is represented using latent variables in the compressed domain (step 202). In a case, the raw image encoder may be a convolutional neural network-based encoder with quantization.

[0078] One or more image signal processing operations, each configured to operate in the compressed domain, may be performed on the latent variables to generate processed latent variables (step 203). The one or more image signal processing operations may be used here to provide any suitable image processing functionality. In a case, for instance, the image signal processing operation(s) may include compressed domain denoising, compressed domain sharpening, compressed domain point spread function (PSF) inversion, compressed domain segmentation, compressed domain motion map estimation and motion compensation, compressed domain motion image registration, or any combination thereof. Using an output image decoder, the processed latent variables may be mapped to an output image in an output color space (step 204). The output image may have any suitable resolution and image format. In a case, the output image may be mapped to a red-green-blue (RGB) color space or a luminance-chrominance (YUV) color space.

[0079] Although FIGURE 2 illustrates an example of a method 200 for image signal processing by using a compressed domain AI, various changes may be made to FIGURE 2. For example, while shown as a series of steps, various steps in FIGURE 2 may overlap, occur in parallel, occur in a different order, or occur any number of times (including zero times).

[0080] FIGURE 3 illustrates an example of an architecture 300 for derivation of a compressed domain for an AI model in accordance with an embodiment of the disclosure. For ease of explanation, the architecture 300 of FIGURE 3 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the architecture 300 may be used by any other suitable device(s) and in any other suitable system(s).

[0081] As shown in FIGURE 3, the architecture 300 may be used for deriving a compressed domain for an AI model using counterpart images, namely a teacher image 301 and a corresponding raw image 302 in this example. Note that the teacher image 301 in FIGURE 3 may conform to any color space, such as an RGB color space, a YUV color space, a hue-saturation-value (HSV) color space, a hue-saturation-lightness (HSL) color space, a hue-saturation-brightness (HSB) color space, or a cyan-magenta-yellow (CMY) color space. As used herein, "teacher image" and "output image" (and similar counterpart terms) should be understood to use the same color space in an embodiment. The teacher image may be referred to as an RGB image. The corresponding raw image 302 may represent a raw data file format specific to a camera or other imaging system and may be, for example, a (quad) Bayer image, a (quad) Tetra image, a hexa-deca image, or an RGBW image. For purposes of this disclosure, while RGB may be used in various examples, any of a number of standardized color spaces, such as those based on a cross-platform color model in which visual perception is balanced against data size, may be used. Raw data generally refers to unprocessed or minimally processed data from an image sensor with (usually) large amounts of potentially redundant data.

[0082] Compared with raw images, RGB images (representative of one possible teacher image color space) contain more semantic information, which makes derivation of a compressed domain easier. Instead of directly designing a compressed domain by training a compression model for raw images, an external teacher model is introduced as shown in FIGURE 3. For example, a teacher image encoder 303 (or a trained teacher image encoder) may take the teacher image 301 as an input and may generate compressed latent variables for a latent manifold 304. The teacher image encoder 303 may be referred to as an RGB image encoder. An output image decoder 305 (or a trained output image decorder) may generate an output image 306 (which utilizes the same color space as the teacher image 301) based on the latent variables within the latent manifold 304. Differences between the teacher image 301 (or the original teacher image) and the output image 306 (or the generated teacher image), such as reconstruction loss, generative adversarial network (GAN) loss, and / or perceptual loss, may be employed in refining the latent variables for the latent manifold 304 (or the compressed latent manifold). As illustrated in FIGURE 3, the architecture 300 may follow an auto-encoder approach. The output image decoder 305 may be referred to as an RGB image decoder.

[0083] Via transfer learning 307, a raw image encoder 308 may be employed to augment the latent variables within the latent manifold 304 to account for latent distance loss. Compared to directly training an encoder and decoder for raw images, this approach is less challenging. The design of the compressed domain (represented by the latent manifold 304) for the raw image 302 is highly supervised and benefits by the teacher image encoder 303 and the output image decoder 305.

[0084] Although FIGURE 3 illustrates one example of an architecture 300 for derivation of a compressed domain for an AI model, various changes may be made to FIGURE 3. For example, any suitable number of images 301 and corresponding raw images 302 may be used here.

[0085] FIGURE 4 illustrates an example of an architecture 400 for a compressed domain for a number of different data formats for an AI model in accordance with an embodiment of the disclosure. For ease of explanation, the architecture 400 of FIGURE 4 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the architecture 400 may be used by any other suitable device(s) and in any other suitable system(s).

[0086] As shown in FIGURE 4, similar to the architecture 300, the architecture 400 may utilize a teacher image encoder 403 and a output image decoder 405 to jointly define a latent manifold 404 (or a compressed latent manifold) using a teacher image 401 (or an input teacher image). Latent variables in the latent manifold 404 may be derived by the teacher image encoding and output image decoding. The latent variables may be refined based on generation of an output image 406 based on the latent variables and assessment of differences between the teacher image 401 and the output image 406.

[0087] Transfer learning 407 may be employed to train a raw image encoder 408 (or a raw data image encoder) that operates on a corresponding raw image 402 (or a raw data image). In a case, the raw image encoder 408 may include multiple encoders for multiple data formats, such as a Bayer image encoder 410 (or a Bayer format image encoder) and a Tetra image encoder 411 (or a Tetra format image encoder). The latent manifold 404 may define a compressed domain that is shared by the various data formats involved, such as RGB, Bayer, and Tetra in the example shown. In this way, downstream ISP tasks may be performed in the same space without extra data conversion. Demosaicing of raw images may be inherently performed by compressing the raw images onto the latent manifold 404 and decoding with the output image decoder 405. An AI model 412 may operate in the compressed domain of the latent manifold 404 and may be used to perform one or more image signal processing (ISP)-related functions.

[0088] Although FIGURE 4 illustrates an example of an architecture 400 for a compressed domain for a number of different data formats for an AI model, various changes may be made to FIGURE 4. For example, any suitable number of images 401 and raw images 402 may be used here. The specific image domains shown here are examples only.

[0089] FIGURE 5 illustrates an example of an architecture 500 for operation of an AI model within a compressed domain in accordance with an embodiment of the disclosure. For ease of explanation, the architecture 500 of FIGURE 5 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the architecture 500 may be used by any other suitable device(s) and in any other suitable system(s).

[0090] As shown in FIGURE 5, similar to the architectures 300 and 400, the architecture 500 may utilize a teacher image encoder 503 (or a trained teacher image encoder) and a output image decoder 505 (or a trained output image encoder) to jointly define a latent manifold 504 (a compressed latent manifold) using a teacher image 501 (or an input teacher image). Latent variables in the latent manifold 504 may be derived by the teacher encoding and output decoding. The latent variables may be refined based on generation of an output image 506 based on the latent variables and assessment of differences between the teacher image 501 and the output image 506.

[0091] Transfer learning 507 may be employed to train a raw image encoder 508 (or a raw data image encoder) that operates on a raw image 502 (or a raw data image). In a case, the raw image encoder 508 may include multiple encoders for multiple data formats, such as a Bayer image encoder 510 and a Tetra image encoder 511. The latent manifold 504 may define a compressed domain that is shared by the various data formats involved, such as RGB, Bayer, and Tetra in the example shown. An AI model 512 may operate in the compressed domain of the latent manifold 504 and may be used to perform one or more ISP-related functions.

[0092] In an embodiment involving compressed domain denoising, a raw image 502 may be denoised directly in the compressed latent space of the latent manifold 504 by designing a network for the AI model 512 that operates on the latent space to perform denoising. In an embodiment involving compressed domain image sharpening, a raw image 502 may be sharpened directly in the compressed latent space of the latent manifold 504 by designing a network for the AI model 512 that operates on the latent space to perform image sharpening. In an embodiment involving compressed domain image sharpening, a raw image 502 may be sharpened directly in the compressed latent space of the latent manifold 504 by designing a network for the AI model 512 that operates on the latent space to perform PSF inversion. In an embodiment involving compressed domain image segmentation, a raw image 502 may be segmented directly in the compressed latent space of the latent manifold 504 by designing a network for the AI model 512 that operates on the latent space to perform segmentation. In an embodiment involving compressed domain motion map estimation, a raw image 502 within a sequence of raw images may be processed directly in the compressed latent space of the latent manifold 504 by designing a network for the AI model 512 that operates on the latent space to perform motion map estimation for the sequence of raw images. In an embodiment involving compressed domain motion image registration, a raw image 502 within a sequence of raw images may be processed directly in the compressed latent space of the latent manifold 504 by designing a network for the AI model 512 that operates on the latent space to perform motion image registration for the sequence of raw images. Note that the same types of operations may occur using the AI model 412 of FIGURE 4.

[0093] Although FIGURE 5 illustrates an example of an architecture 500 for operation of an AI model within a compressed domain, various changes may be made to FIGURE 5. For example, any suitable number of teacher images 501 and raw images 502 may be used here. The specific image domains shown here are examples only.

[0094] FIGURE 6 illustrates an example of an AI model 600 for efficient raw domain compressed image data ISP in accordance with an embodiment of the disclosure. For ease of explanation, the AI model 600 of FIGURE 6 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the AI model 600 may be used by any other suitable device(s) and in any other suitable system(s).

[0095] As shown in FIGURE 6, the model 600 may include an encoder 601, a quantization layer 602, and a decoder 603. In an embodiment, the encoder 601 may include a plurality of feed-forward convolutional neural network (CNN) layers 604-607, each of which can learn features via kernel optimization by applying a cascaded convolution (cross-correlation) kernel. In an embodiment, the CNN layers 604-607 of the encoder 601 may be of kernel size 5×5 and have a stride of two. Outputs of the CNN layer 605 and the CNN layer 607 may be passed to local or sliding windowed attention layers 608 and 609, respectively, each of which uses a fixed window size. The model 600 is adapted for efficient compression of raw data by pruning and modifying blocks. Within the encoder 601, model pruning may be performed by removing one or more parameters in a manner substantially maintaining accuracy while reducing the model size and / or increasing efficiency.

[0096] The quantization layer 602 may compress the output of the encoder 601, such as by reducing the precision of weights (and optionally also bias and activation functions) so that the model requires less memory and may run on resource-constrained devices. The decoder 603 may be essentially an inverse of the encoder 601 and may include a plurality of transposed CNN layers 614-617 and two windowed attention layers 618 and 619. In an embodiment, the transposed CNN layers 614-617 of the decoder 603 are of kernel size 5×5 and have a stride of two.

[0097] Although FIGURE 6 illustrates an example of an AI model 600 for efficient raw domain compressed image data ISP, various changes may be made to FIGURE 6. For example, the AI model 600 may include any suitable number of layers in the encoder 601 and the decoder 603.

[0098] FIGURE 7 illustrates an example of an architecture 700 for progressive compression-decompression for an AI model operating in the compressed domain in accordance with an embodiment of the disclosure. For ease of explanation, the architecture 700 of FIGURE 7 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the architecture 700 may be used by any other suitable device(s) and in any other suitable system(s).

[0099] As shown in FIGURE 7, similar to the architectures of FIGURES 3, 4, and 5, the architecture 700 of FIGURE 7 utilizes a teacher image encoder 703 (or a trained teacher data encoder) and a full-resolution output decoder 705 (or a trained full-resolution output data decoder) to jointly define a latent manifold 704 (or a compressed latent manifold) using a teacher image 701(or an input teacher image). However, the architecture 700 may utilize a low-resolution output decoder 725 (or a trained low-resolution output data decoder) for defining the latent manifold 704. The inclusion of the low-resolution output decoder 725 may allow generation of an output image 726 with lower resolution (such as a "preview" image) than the teacher image 706 generated by the full-resolution output decoder 705. In a case, this may be accomplished using only a proportion of the latent variables in the latent manifold 704. Downstream ISP tasks that do not require full-resolution image processing can therefore benefit from this approach. For example, motion map estimation or tone mapping analysis may save computations and memory usage with the progressive compression-decompression of FIGURE 7.

[0100] Although FIGURE 7 illustrates an example of an architecture 700 for progressive compression-decompression for an AI model operating in the compressed domain, various changes may be made to FIGURE 7. For example, any suitable number of teacher images 701 may be used here.

[0101] FIGURES 8A through 8C comparatively illustrate example compression and inherent demosaicing for raw data when using a compressed domain AI for image signal processing in accordance with an embodiment of the disclosure. FIGURE 8A illustrates a raw (Tetra) image (or image data) input to a raw image encoder (or a raw data encoder) and a corresponding image (or image data) output from an output image decoder (or an output data decoder). By comparison, FIGURE 8B illustrates a teacher image (or teacher image data or an RGB image) input to a teacher image encoder (or a teacher data encoder or an RGB image encoder) and a corresponding image (or image data) output from an output image decoder (or an output data decoder). FIGURE 8C illustrates a ground truth image (or ground truth image data).

[0102] FIGURES 9A through 9C comparatively illustrate example compression and inherent demosaicing for raw data when using a compressed domain AI for image signal processing using an enlargement of a portion of the image content from FIGURES 8A through 8C in accordance with an embodiment of the disclosure. FIGURE 9A illustrates the output for a raw (such as Tetra) image (or raw image data) input to a raw image encoder (or a raw data encoder) without a teacher model and transfer learning from a teacher image encoder (or a trained teacher data encoder) and a corresponding output image decoder (or a corresponding trained output data decoder) for image data output. By comparison, FIGURE 9B illustrates the output for the same raw image input to a raw image encoder with a teacher model and transfer learning. FIGURE 9C illustrates ground truth image.

[0103] FIGURES 10A through 10C comparatively illustrate an example the result of raw image denoising in the compressed latent domain for image signal processing in accordance with an embodiment of the disclosure. FIGURE 10A illustrates the output image for raw image data compression, compressed domain denoising, and decompression of noisy raw (Tetra) image raw data. By comparison, FIGURE 10B illustrates the output image for the raw image data compression and decompression of the same noisy raw image raw data without compressed domain denoising. FIGURE 10C illustrates ground truth image data. FIGURE 10A shows the result of raw image denoising in the compressed latent domain, where noisy latent variables are input into a denoising module, which in turn outputs clean latent variables.

[0104] FIGURES 11A through 11C comparatively illustrate an example of the result of multi-frame blending in the compressed domain for image signal processing in accordance with an embodiment of the disclosure. FIGURE 11A illustrates multi-frame blending in the compressed domain. By comparison, FIGURE 11B illustrates a single frame in the compressed domain. FIGURE 11C illustrates ground truth image data.

[0105] FIGURES 12A and 12B illustrate an example of the result of progressive encoding and decoding in accordance with this disclosure. FIGURE 12B is the full resolution output image, while FIGURE 12A is a quarter resolution output image from one-quarter of the most significant latent space variables within a latent manifold.

[0106] Although FIGURES 8A through 12B illustrate an example of results of various image processing operations, various changes may be made to FIGURES 8A through 12B. For example, the actual contents of images being processed can vary widely. Also, the specific results that are obtained can vary depending on the implementation.

[0107] FIGURE 13 illustrates an example of an AI model 1300 for efficient raw domain compressed image data ISP in accordance with an embodiment of the disclosure. For ease of explanation, the AI model 1300 of FIGURE 13 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the AI model 1300 may be used by any other suitable device(s) and in any other suitable system(s).

[0108] As shown in FIGURE 13, the model 1300 may be similar to the model 600 of FIGURE 6. As with the model 600, the model 1300 may include an encoder 1301, a quantization layer 1302, and a decoder 1303. The encoder 1301 may include CNN layers 1304-1307 that correspond to the CNN layers 604-607 and windowed attention layers 1308-1309 that correspond to the windowed attention layers 608-609. The quantization layer 1302 may be analogous to the quantization layer 602. The decoder 1303 may include transposed CNN layers 1314-1317 that correspond to the transposed CNN layers 614-617 and windowed attention layers 1318-1319 that correspond to the windowed attention layers 608-609.

[0109] The decoder 1303 in FIGURE 13 may include a generalized divisive normalization (GDN) layer 1321 between the CNN layer 1304 and the CNN layer 1305, a GDN layer 1322 between the CNN layer 1305 and the windowed attention layer 1308, and a GDN layer 1323 between the CNN layer 1306 and the CNN layer 1307. Each of the GDN layers 1321-1323 may introduce a nonlinearity that improves the efficiency of the transforms by using reparameterization. The decoder 1303 may include an inverse GDN layer 1331 between the transposed CNN layer 1314 and the transposed CNN layer 1315, an inverse GDN layer 1332 between the transposed CNN layer 1315 and the windowed attention layer 1318, and an inverse GDN layer 1333 between the transposed CNN layer 1316 and the transposed CNN layer 1317. Each of the inverse GDN layers 1331-1333 may perform the opposite function of the GDN layer 1321-1323, respectively. The GDN layers 1321-1323 and the inverse GDN layers 1331-1333 may improve the performance of the model 1300 relative to the model size.

[0110] Although FIGURE 13 illustrates another example of an AI model 1300 for efficient raw domain compressed image data ISP, various changes may be made to FIGURE 13. For example, the AI model 1300 may include any suitable number of layers in the encoder 1301 and the decoder 1303.

[0111] FIGURE 14 illustrates an example of an architecture 1400 for generalized progressive compression-decompression for an AI model operating in the compressed domain in accordance with an embodiment of the disclosure. For ease of explanation, the architecture 1400 of FIGURE 14 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the architecture 1400 may be used by any other suitable device(s) and in any other suitable system(s).

[0112] As shown in FIGURE 14, similar to the architecture 700 of FIGURE 7, the architecture 1400 may utilize an encoder 1403 (or a trained data encoder, or a progressive data encoder) for defining a latent manifold 1404 (or a compressed latent manifold) based on an input image 1401. The architecture 1400 may utilize a trained full-resolution data decoder 1405 (or model 2 or a full-resolution data (or image) decoder) and a trained low-resolution data decoder 1425(or model 1 or a low-resolution data (or image) decoder). However, the trained full-resolution data decoder 1405 generating a full-resolution image 1406 and the trained low-resolution data decoder 1425 generating a low-resolution image 1426 need not be models for performing the same type of image analysis or image processing. In a case, the image processing pipeline may include both pixel-level processing tasks, such as denoising and demosaicing, and global-level analysis tasks, such as tone-mapping gain estimation and motion estimation. Not all analysis tasks need full-resolution images.

[0113] With the encoder 1403, the most significant bits of the compressed data may be decoded into a low-resolution image 1426 using the trained low-resolution data decoder 1425, while a full-resolution image 1406 may be recovered using the trained full-resolution data decoder 1405. The trained low-resolution data decoder 1425 may be referred to as a lightweight low-resolution data decoder. For analysis tasks such as registration, the most-significant bits of the compressed data may be used to save computation and memory. For other tasks, the full compressed data may be used. The trained low-resolution data decoder 1425 and the trained full-resolution data decoder 1405 may be different AI models ("Model 1" and "Model 2", respectively) performing different image analysis or image processing tasks. For example, "Model 1" may perform a global-level processing task such as image registration, while "Model 2" performs a pixel-level processing task such as denoising.

[0114] Although FIGURE 14 illustrates an example of an architecture 1400 for generalized progressive compression-decompression for an AI model operating in the compressed domain, various changes may be made to FIGURE 14. For example, any suitable number of images 1401 may be used here.

[0115] FIGURE 15 illustrates an example of an architecture 1500 for derivation of a compressed domain for an AI model utilizing generalized progressive compression-decompression in accordance with an embodiment of the disclosure. For ease of explanation, the architecture 1500 of FIGURE 15 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the architecture 1500 may be used by any other suitable device(s) and in any other suitable system(s).

[0116] As shown in FIGURE 15, similar to the architecture 300 of FIGURE 3, the architecture 1500 may utilize a teacher image encoder 1503 (or a trained teacher image encoder) with a teacher image 1501 as an input and may generate compressed latent variables for a latent manifold 1504. A full-resolution output image decoder 1505 (or a trained full-resolution output image decoder) for generating a full-resolution output image 1506 based on the latent variables within the latent manifold 1504 may be provided, together with a low-resolution output image decoder 1525 (or a trained low-resolution output image decoder) for generating a low-resolution output image 1526. Differences between the teacher image 1501 (or the original teacher image) and both of the generated full-resolution output image 1506 and the generated low-resolution output image 1526 may be employed in refining the latent variables for the latent manifold 1504 (or the compressed latent manifold). In a case, the differences may encompass RGB reconstruction loss, GAN loss, and / or perceptual loss.

[0117] Via transfer learning 1507, the raw image encoder 1508 may be employed to augment the latent variables within the latent manifold 1504. The raw image encoder 1508 may be referred to as a full-resolution raw image encoder. This may help to account for latent distance loss. The latent manifold 1504 may be employed based on a raw data image 1502, such as by different AI models performing different image analysis or image processing tasks.

[0118] Although FIGURE 15 illustrates an example of an architecture 1500 for derivation of a compressed domain for an AI model utilizing generalized progressive compression-decompression, various changes may be made to FIGURE 15. For example, any suitable number of teacher images 1501 and corresponding raw images 1502 may be used here.

[0119] FIGURE 16 illustrates an example of an architecture 1600 for generalized progressive compression-decompression for an AI model in accordance with an embodiment of the disclosure. For ease of explanation, the architecture 1600 of FIGURE 16 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the architecture 1600 may be used by any other suitable device(s) and in any other suitable system(s).

[0120] As shown in FIGURE 16, similar to the architecture 1400 of FIGURE 14, the architecture 1600 may utilize an encoder 1603 (or a trained data encoder) for defining a compressed latent manifold 1604 based on an input image 1601. Although not explicitly shown, the encoder 1603 may use the same layers as the encoder 1301 in FIGURE 13, including GDN layers. The architecture 1600 may utilize a full-resolution data decoder 1605 for generating a full-resolution image 1606 and a low-resolution data decoder 1625 for generating a low-resolution image 1626. Both the full-resolution data decoder 1605 and the low-resolution data decoder 1625 may include inverse GDN layers. In the example shown, the full-resolution data decoder 1605 may use the same layers as the decoder 1303 in FIGURE 13, while the low-resolution data decoder 1625 may use one less transposed CNN layer and one less inverse GDN layer than the full-resolution data decoder 1605.

[0121] Although FIGURE 16 illustrates an example of an architecture 1600 for generalized progressive compression-decompression for an AI model, various changes may be made to FIGURE 16. For example, any suitable number of images 1601 may be used here.

[0122] FIGURES 17A through 17C comparatively illustrate example compression and inherent demosaicing for raw data when using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with an embodiment of the disclosure. FIGURE 17A illustrates a low-resolution output image for raw (Tetra) image data input to a full-resolution raw data encoder without a teacher model and transfer learning from a trained teacher image encoder (or a trained RGB data encoder) and corresponding trained raw image decoder (or corresponding trained RGB data decoder) for raw image data output (or RGB image data output). By comparison, FIGURE 17B illustrates the low-resolution output image for the same raw image data input to a raw image encoder with a teacher model and transfer learning. FIGURE 17C illustrates ground truth image. The compression ratio for low-resolution image processing in the example shown is five.

[0123] FIGURES 18A through 18C comparatively illustrate example raw image denoising using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with an embodiment of the disclosure. FIGURE 18A illustrates an output image for raw image data compression, compressed domain denoising, and low-resolution decompression of noisy raw (Tetra) image raw data. By comparison, FIGURE 18B illustrates an output image for the raw image data compression and low-resolution decompression of the same noisy raw image raw data without compressed domain denoising. FIGURE 18C illustrates ground truth image. The compression ratio for low-resolution image processing in the example shown is five.

[0124] FIGURES 19A through 19C comparatively illustrate an example of multi-frame blending using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with this disclosure. FIGURE 19B illustrates multi-frame blending in the compressed domain. By comparison, FIGURE 19A illustrates a single frame in the compressed domain. FIGURE 19C illustrates ground truth image. The compression ratio for low-resolution image processing in the example shown is five. The compressed domain multi-frame blending effectively reduces noise while maintaining details.

[0125] FIGURES 20A through 20C comparatively illustrate example image registration using a low-resolution decoder for a compressed domain AI performing image signal processing in accordance with an embodiment of the disclosure. FIGURE 20A illustrates the original image, FIGURE 20B illustrates the warped image, and FIGURE 20C illustrates the target image. The warped image of FIGURE 20B closely may matche the target of FIGURE 20C, indicating that compressed domain registration parameter estimation is accurate. The compression ratio for low-resolution image processing in the example shown is five.

[0126] Although FIGURES 17A through 20C illustrate an example of results of various image processing operations, various changes may be made to FIGURES 17A through 20C. For example, the actual contents of images being processed may vary widely. The specific results that are obtained may vary depending on the implementation.

[0127] FIGURE 21 illustrates a multi-frame ISP pipeline 2100. In an embodiment, the pipeline 2100 may operate on multiple images 2101 and may include a sequence of various components 2102-2104 that successively operate on image data to produce an output image 2105. Any of the ISP tasks performed by the components 2102-2104 may be performed using AI-based implementations.

[0128] FIGURE 22 illustrates an example of a multi-frame ISP pipeline 2200. Here, the pipeline 2200 may operate on multiple images 2201. An image registration component 2202 may align the multiple images 2201. A demosaicing and blending component 2203 may estimate a color image that most closely resembles the original image represented by the multiple images 2201 and blend information from the multiple images 2201 together to form a single image 2206. A tone mapping component 2204 may map colors in the single image 2206 to another set of colors to approximate the appearance of a high dynamic range (HDR) image, which corresponds to an output image 2205.

[0129] There is an increasing need for high-megapixel imaging systems (such as 50 megapixels and up) using multi-frame imaging pipelines of the type depicted in FIGURES 21 and 22. These pipelines may be used for imaging to achieve noise-reduction and HDR, in light of the following problems. First, memory and computing constraints may limit the number of images that may be used for high-megapixel pipelines. Second, conventional compression methods may be not directly designed for compressed domain ISP downstream tasks, such as demosaicing and / or denoising. Reducing the memory footprint of high-megapixel pipelines with conventional compression often may necessitate repeatedly de-compressing, processing, and compressing data, resulting in high compute / memory overhead. Third, the compression ratio for conventional compression methods may be typically image content-dependent. As a result, reducing the memory footprint of high-megapixel pipelines with conventional compression may become challenging.

[0130] FIGURE 23 illustrates a pipeline 2300 for the ISP sequence of FIGURE 22 in the context of a compute engine 2310 and memory 2311. In the pipeline 2300, images 2301 may be loaded into a memory 2311, may read from the memory 2311 for processing, may be processed by the compute engine 2310, and may be written to the memory 2311 again for subsequent processing. In FIGURE 23, the pipeline 2300 may require an initial memory write, three memory read-write pairs (one each for the image registration component 2302, the demosaicing and blending component 2303, and the tone mapping component 2304), and a final memory read for the output image 2305. The memory read and write requirements may limit the size (N) of the images and the number of the images that may be used in the pipeline 2300. High-resolution cameras may exacerbate the problem.

[0131] FIGURE 24 illustrates a pipeline 2400 for the ISP sequence of FIGURE 22 in the context of a compute engine 2410 and memory 2411 and assuming a specified data compression ratio. Memory compression using conventional compression methods may be used to reduce memory consumption. However, ISP downstream tasks (such as demosaicing and / or denoising) may be typically not designed to operate with such compressed data. Therefore, in the pipeline 2400, input images 2401 may be loaded into a memory 2411 and may be processed by the image registration component 2402, the demosaicing and blending component 2403, and the tone mapping component 2404 to produce an output image 2405. The input images 2401 may be subject to compression 2420 when initially written to the memory 2411. Performing the ISP downstream tasks (image registration, demosaicing and blending, and tone mapping) may require decompression 2421, 2423, and 2425 of the input images 2401 to full size, performance of the corresponding full-size image processing, compression 2422, 2424, and 2426 of the output from each of the image registration component 2402, the demosaicing and blending component 2403, and the tone mapping component 2404 when the processed full-size data is written to memory, and decompression 2427 of the output image 2405. Repeated decompression, processing, and compression steps generally result in high compute / memory overhead.

[0132] Memory compression using conventional compression techniques may be designed for standardized color space images rather than raw format images, particularly those with certain color filter array (CFA) patterns such as Bayer, Tetra, etc. Raw data processing may be useful or important in modern ISPs, such as for 50-megapixel cameras equipped with CFAs. Applying such conventional compression techniques to raw data results in loss of detail and the creation of other image artifacts.

[0133] With increasing resolution, memory requirements for multi-frame imaging pipelines may become a bottleneck for downstream tasks of such pipelines. That problem may persist even with AI-based implementations of downstream tasks of the multi-frame imaging pipelines. The approach to address these issues described above may be to compress high-resolution images to a compressed domain with smaller size and perform the downstream tasks, such as demosaicing and denoising, in the compressed domain. As noted above, designing such an effective and general compression domain for downstream tasks remains challenging.

[0134] A compression technique that is well-suited for all image formats (such as both RGB and raw) may compress high-resolution images to a compressed domain (referred to here as a latent space or latent manifold) where the images may be represented with a reduced set of features (referred to here as latent variables), resulting in a smaller-size representation. An AI-based encoder that compresses high-resolution images from an image space to a compressed domain may be employed, together with a novel AI-based decoder that decompresses images from the compressed domain to the image space. AI-based ISP pipeline tasks that are suited for the compressed domain may be employed, providing a way to perform one, some, or all of the downstream ISP pipeline tasks in the compressed domain. As a result, a system may be suited for the compressed domain, may be compatible with one or more downstream ISP tasks (even for high-megapixel pipelines), and may reduce or minimize image losses and artifacts. Use cases may be enabled that are not currently possible due to a lack of memory.

[0135] In an embodiment of the disclosure, a pipeline may utilize AI to compress data and perform AI-based processing on compressed data. For AI-based compression and processing according to an embodiment of the disclosure, the pipeline may use AI encoder networks to compress images to an encoded space that has a smaller size than the original image(s), thereby providing memory savings. AI-based processing may be performed on the compressed image(s) to output one or more processed compressed domain images. AI decoder networks may be used to decode the compressed image(s) into one or more uncompressed images. For multi-frame images, the pipeline may utilize AI to compress the multi-frame images and may perform AI-based processing on the compressed images. For multi-task processing, the pipeline may utilize AI to compress sensor data and perform AI-based processing on the compressed sensor data to complete various downstream tasks.

[0136] FIGURE 25 illustrates an example of a pipeline 2500 utilizing AI to compress data and performing AI-based processing on the compressed data in accordance with an embodiment of the disclosure. For ease of explanation, the pipeline 2500 of FIGURE 25 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the pipeline 2500 may be used by any other suitable device(s) and in any other suitable system(s).

[0137] As shown in FIGURE 25, the pipeline 2500 may use AI encoder networks to compress images 2501 to an encoded space so that each encoded image has a smaller size than the corresponding original image, thereby providing memory savings. AI-based processing may be used on the compressed images to output processed compressed domain images, and AI decoder networks may be used to decode the compressed images into uncompressed images 2505. In the context of a compute engine 2510 and memory 2511, the pipeline 2500 may include a first model 2502 ("Model 1", such as an image registration component), a second model 2503 ("Model 2", such as a demosaicing and blending component), and a third model 2504 ("Model 3", such as a tone mapping component). An AI encoder 2551 may execute on the compute engine 2510, writing compressed domain image data (the latent variables for a latent manifold) to the memory 2511. The first model 2502, the second model 2503, and a third model 2504 may operate on the compressed domain image data. An AI decoder 2552 may execute on the compute engine 2510, reading processed compressed domain image data from the memory 2511 and decoding for output as the uncompressed images 2505.

[0138] Within the pipeline 2500, if a downstream task uses pixel-wise operations (such as for a denoising operation), a decoder may handle that operation. However, if the downstream task uses global operations (such as for a registration operation), a compressed domain model may handle the operation. The pipeline 2500 may process multi-frame data and, for multi-task operations, may perform multiple downstream image processing tasks using AI-based processing on the compressed domain image data.

[0139] Although FIGURE 25 illustrates an example of a pipeline 2500 utilizing AI to compress data and performing AI-based processing on the compressed data, various changes may be made to FIGURE 25. For example, while three models are shown here, the pipeline 2500 may support the use of any suitable number of models.

[0140] FIGURE 26 illustrates an example of a pipeline 2600 utilizing AI to compress data and performing AI-based processing on the compressed data in accordance with an embodiment of the disclosure. For ease of explanation, the pipeline 2600 of FIGURE 26 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the pipeline 2600 may be used by any other suitable device(s) and in any other suitable system(s).

[0141] As shown in FIGURE 26, similar to the pipeline 2500 of FIGURE 25, the pipeline 2600 may use AI encoder networks 2651 to compress images 2601 and AI decoder networks 2652 to decode compressed images into uncompressed images 2605. However, the AI encoder networks 2651 and the AI decoder networks 2652 need not execute on the same compute engine 2610 as the AI model(s) processing the compressed domain image data. Instead, as shown in FIGURE 26, the AI encoder networks 2651 may execute separately from the compute engine 2610 and pass the compressed domain image data to the compute engine 2610. The compute engine 2610 may store the received compressed domain image data in memory 2611. The compute engine 2610 may retrieve portions of the stored compressed domain data as needed or desired for processing by a first model 2602 (such as an image registration component) and a second model 2603 (such as a blending component). Processed compressed domain image data may be passed to the AI decoder networks 2652 for output.

[0142] AI-based data compression and AI-based processing on the compressed data may exploit progressive compression and processing. With progressive compression and processing, the pipeline 2600 may utilize the AI encoder networks 2651 to perform progressive AI compression (such as when the compressed data is ordered from most significant bits to least significant bits), and AI-based processing may be performed (such as by the first model 2602 performing image registration) on only the most significant bits of the compressed data to efficiently perform downstream analysis tasks that do not need full-resolution data. With progressive AI compression, the most-significant bits of the compressed data may be used to get low-resolution images that may be used for performing tasks on the compressed domain image data that do not require full-resolution data, such as registration, tone-map gain computation, etc.

[0143] Although FIGURE 26 illustrates an example of a pipeline 2600 utilizing AI to compress data and performing AI-based processing on the compressed data, various changes may be made to FIGURE 26. For example, while two models are shown here, the pipeline 2600 may support the use of any suitable number of models.

[0144] FIGURE 27 illustrates an example of a compressed domain AI multi-frame multi-task pipeline 2700 in accordance with an embodiment of the disclosure. For ease of explanation, the pipeline 2700 of FIGURE 27 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the pipeline 2700 may be used by any other suitable device(s) and in any other suitable system(s).

[0145] As shown in FIGURE 27, a stream of input frames 2701 may be received by an application processor 2702 that performs AI compression 2703 on the stream of input frames 2701 and outputs a stream of compressed data 2704. A first compressed domain AI model 2705 may receive the compressed data 2704 and perform an AI operation 2706 (such as image registration), outputting a processed stream of compressed data 2707. A compressed domain image processor 2708 may perform an image processing function (such as blending 2709) on the stream of processed stream of compressed data 2707. The compressed domain image processor 2708 may output a single frame of compressed data 2710. A second compressed domain AI model 2711 may receive the single frame of compressed data 2710 and perform an AI operation 2712, outputting a processed single frame of compressed data 2713. The processed single frame of compressed data 2713 may undergo AI decoding 2714 to produce an output image 2715 (or output).

[0146] Although FIGURE 27 illustrates an example of a compressed domain AI multi-frame multi-task pipeline 2700, various changes may be made to FIGURE 27. For example, while three compressed domain operations are shown here, the pipeline 2700 may support the use of any suitable number of compressed domain operations. Also, any other or additional compressed domain operations may be performed here.

[0147] FIGURE 28 illustrates an example of a pipeline 2800 utilizing progressive AI compression and using most significant bits to efficiently perform downstream tasks that do not need full-resolution data in accordance with an embodiment of the disclosure. For ease of explanation, the pipeline 2800 of FIGURE 28 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the pipeline 2800 may be used by any other suitable device(s) and in any other suitable system(s).

[0148] As shown in FIGURE 28, similar to the pipeline 2700 of FIGURE 27, the pipeline 2800 may receive a stream of input frames 2801 at an application processor 2802 that performs AI compression 2803 on those input frames 2801 and outputs a stream of compressed data 2804. The AI compression 2803 may be progressive (such as when the compressed data is ordered from most significant bits 2805 to least significant bits 2806). A first compressed domain AI model 2807 may receive the compressed data 2804 and perform an AI operation 2808 that operates on all of the compressed data 2804, outputting a processed stream of compressed data 2809. A second compressed domain AI model 2810 may receive the compressed data 2809 and perform an AI operation 2811 that operates on only the most significant bits of the compressed data 2809, efficiently performing an analysis task that does not need full-resolution data. The second compressed domain AI model 2810 may output all of the compressed data 2812 but possibly with only the most significant bits modified. Additional processing on the compressed data 2812 may be performed within the pipeline 2800, producing compressed data 2813 that undergoes AI decoding 2814 to produce an RGB output image 2815.

[0149] In a pipeline, there are analysis steps (such as registration) that do not require full-resolution images. For these cases, operating on the full compressed data may be unnecessary and computationally expensive. Instead, with a progressive encoder (such as AI compression 2803 performed by the application processor 2802 in the pipeline 2800), the most significant bits of the compressed data may be decoded into a low-resolution image using a separate lightweight low-resolution decoder (not shown), while the full-resolution image may be recovered using the full decoder (such as AI decoding 2814). For analysis tasks such as registration, only the most-significant bits of the compressed data might be used. For other tasks, the full compressed data may be used.

[0150] Although FIGURE 28 illustrates an example of a pipeline 2800 utilizing progressive AI compression and using most significant bits to efficiently perform downstream tasks that do not need full-resolution data, various changes may be made to FIGURE 28. For example, while two compressed domain operations are shown here, the pipeline 2800 may support the use of any suitable number of compressed domain operations. Also, any other or additional compressed domain operations may be performed here.

[0151] FIGURE 29 illustrates an example of a pipeline 2900 utilizing progressive AI compression and using most significant bits to efficiently perform downstream tasks that do not need full-resolution data in accordance with an embodiment of the disclosure. For ease of explanation, the pipeline 2900 of FIGURE 29 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the pipeline 2900 may be used by any other suitable device(s) and in any other suitable system(s).

[0152] As shown in FIGURE 29, similar to the pipelines of FIGURES 27 and 28, the pipeline 2900 may receive a stream of input frames 2901 at an application processor 2902 that performs AI compression 2903 on those input frames 2901 and outputs a stream of compressed data 2904. The AI compression 2903 is progressive (such as when the compressed data is ordered from most significant bits to least significant bits). Unlike the sequential AI operations 2808 and 2811 in the pipeline 2800, AI operations may be performed in tandem in the pipeline 2900. In the embodiment, the most significant bits 2904 of the compressed data 2905 may be received by a first compressed domain AI model 2906 that performs an AI operation 2907 such as registration to obtain motion vectors 2908. The full set of the compressed data 2904 may be received by a second compressed domain AI model 2909 that performs an AI operation 2910 such as a warp-blend operation. The outputs of the AI operation 2907 and the AI operation 2910 may be combined to generate processed compressed data 2911, which undergoes AI decoding 2912 to produce an RGB output image 2913.

[0153] Although FIGURE 29 illustrates an example of a pipeline 2900 utilizing progressive AI compression and using most significant bits to efficiently perform downstream tasks that do not need full-resolution data, various changes may be made to FIGURE 29. For example, while two compressed domain operations are shown here, the pipeline 2900 may support the use of any suitable number of compressed domain operations. Also, any other or additional compressed domain operations may be performed here.

[0154] FIGURE 30 illustrates an example of a pipeline 3000 utilizing AI to compress multi-frame data using temporal redundancy between frames in accordance with an embodiment of the disclosure. For ease of explanation, the pipeline 3000 of FIGURE 30 is described as being used by the electronic device 101 in the network configuration 100 of FIGURE 1. However, the pipeline 3000 may be used by any other suitable device(s) and in any other suitable system(s).

[0155] In a case, AI-based data compression and AI-based processing on the compressed data may exploit temporal redundancy during encoding. For example, with temporal redundancy-based encoding, the pipeline 3000 may utilize AI to compress multi-frame data 3001 by exploiting the temporal redundancy between the frames of the multi-frame data 3001 and performs AI-based processing on the compressed data. The pipeline 3000 of FIGURE 30 may receive the stream of input frames for the multi-frame data 3001 at an application processor 3002 that performs AI compression 3003 using temporal redundancy between those input frames for the multi-frame data 3001. The application processor 3002 may output a stream of compressed data 3004. The stream of compressed data 3004 may be subject to subsequent processing 3005 to generate a processed single frame of compressed data 3006, which undergoes AI decoding 3007 to produce an output image 3008.

[0156] Although FIGURE 30 illustrates an example of a pipeline 3000 utilizing AI to compress multi-frame data using temporal redundancy between frames, various changes may be made to FIGURE 30. For example, the processing 3005 may include any suitable number and types of compressed domain operations.

[0157] FIGURES 31A through 31C comparatively illustrate an example of a result of multi-frame blending in a compressed domain for AI-based image signal processing in accordance with an embodiment of the disclosure. FIGURE 31A illustrates the effect of multi-frame blending in the compressed domain in which noisy Tetra raw frames are compressed (such as using AI-based compression), multi-frame blending occurs in the compressed domain, and AI decoding to RGB is performed. By contrast, FIGURE 31B illustrates the effect of AI-based compression of noisy Tetra raw frames followed by AI decoding to RGB without multi-frame blending in the compressed domain. FIGURE 31C illustrates paired RGB ground truth obtained by a digital single lens reflex (DSLR) camera by shifting the sensor to collect all three channels at all pixel locations. A compression ratio of five, producing compressed data with five times smaller memory footprint that the raw Tetra input data, is employed with five frames.

[0158] FIGURES 32A through 32D comparatively illustrate an example of a result of compressed domain for AI denoising in accordance with an embodiment of the disclosure. FIGURES 32A and 32B illustrates the effect of AI denoising in the compressed domain in which noisy Tetra raw frames are compressed (such as using AI-based compression), AI denoising occurs in the compressed domain, and AI decoding to RGB is performed. By contrast, FIGURE 32C illustrates the effect of AI-based compression of noisy Tetra raw frames followed by AI decoding to RGB without AI denoising in the compressed domain. FIGURE 32D illustrates paired RGB ground truth obtained by a DSLR camera by shifting the sensor to collect all three channels at all pixel locations. A compression ratio of five is employed in FIGURE 32A, and a compression ratio of 7.5 is employed in FIGURE 32D.

[0159] FIGURES 33A through 33C comparatively illustrate an example of a result of compressed domain for AI registration in accordance with an embodiment of the disclosure. FIGURES 33A and 33C illustrate reference and input frames, respectively. FIGURE 33B illustrates the effect of compressed domain AI registration in which noisy Tetra raw frames are compressed (such as using AI-based compression), AI occurs in the compressed domain, and AI decoding to RGB is performed. Warping is performed via registration parameters obtained from AI registration.

[0160] Although FIGURES 30A through 33B illustrate an example of the results of various image processing operations, various changes may be made to FIGURES 30A through 33B. For example, the actual contents of images being processed may vary widely. Also, the results that are obtained may vary depending on the implementation.

[0161] It should be noted that the functions shown in the figures or described above may be implemented in an electronic device 101, 102, 104, server 106, or other device(s) in any suitable manner. For example, in an embodiment, at least some of the functions shown in the figures or described above may be implemented or supported using one or more software applications or other software instructions that are executed by the processor 120 of the electronic device 101, 102, 104, server 106, or other device(s). In an embodiment, at least some of the functions shown in the figures or described above may be implemented or supported using dedicated hardware components. In general, the functions shown in the figures or described above may be performed using any suitable hardware or any suitable combination of hardware and software / firmware instructions. Also, the functions shown in the figures or described above may be performed by a single device or by multiple devices.

[0162] According to an embodiment of the disclosure, the output color space is a red-green-blue (RGB) color space and the output image (306) is an RGB image, or the output color space is a luminance-chrominance (YUV) color space and the output image (306) is a YUV image.

[0163] According to an embodiment of the disclosure, the raw image (302) comprises a Tetra image, a Bayer image, a hexa-deca image, or a red-green-blue-white (RGBW) image.

[0164] According to an embodiment of the disclosure, the raw image encoder (308) comprises a convolutional neural network-based encoder with quantization or the output image decoder (305) comprises a convolutional neural network.

[0165] According to an embodiment of the disclosure, the one or more image signal processing operations comprise at least one of: compressed domain denoising, compressed domain sharpening, compressed domain point spread function (PSF) inversion, compressed domain segmentation, compressed domain motion map estimation and motion compensation, or compressed domain motion image registration.

[0166] According to an embodiment of the disclosure, the raw image encoder (308) comprises a trained machine learning model subjected to a training scheme comprising: training a teacher image encoder (303) to map input images into the compressed domain, wherein the training yields trained network weights for the teacher image encoder (303); initializing network weights of the raw image encoder (308) from the trained network weights for the teacher image encoder (303); and constraining outputs of the raw image encoder (303) to be close to outputs of the teacher image encoder (303) via a mean squared error loss in the compressed domain.

[0167] According to an embodiment of the disclosure, the teacher image encoder (303) comprises a red-green-blue (RGB) image encoder and the training scheme comprises training the teacher image encoder (303) to map RGB images into the compressed domain.

[0168] According to an embodiment of the disclosure, the teacher image encoder comprises a luminance-chrominance (YUV) image encoder and the training scheme comprises training the teacher image encoder to map YUV images into the compressed domain.

[0169] According to an embodiment of the disclosure, the mapping the raw image to the compressed domain comprises: mapping a multi-frame stream of data including the raw image to the compressed domain based on (i) a first frame within the multi-frame stream of data and (ii) at least one of: spatial redundancy between the first frame and a remainder of frames within the multi-frame stream of data or the spatial redundancy between the first frame and the remainder of frames and temporal redundancy between consecutive frames within the multi-frame stream of data, and the first frame includes the raw image.

[0170] According to an embodiment of the disclosure, the mapping the raw image to the compressed domain comprises ordering data from most significant bits to least significant bit or the performing the one or more image signal processing operations comprises processing only the most significant bits of data in the compressed domain to perform at least one of the one or more image signal processing operations that does not require full-resolution data.

[0171] According to an embodiment of the disclosure, the teacher image encoder is a red-green-blue (RGB) image encoder or a luminance-chrominance (YUV) image encoder and the training scheme comprises training the teacher image encoder to map RGB images or YUV images into the compressed domain.

[0172] According to an embodiment of the disclosure, to map the raw image to the compressed domain, the at least one processing device is configured to map a multi-frame stream of data including the raw image to the compressed domain based on (i) a first frame within the multi-frame stream of data and (ii) at least one of: spatial redundancy between the first frame and a remainder of frames within the multi-frame stream of data, or the spatial redundancy between the first frame and the remainder of frames and temporal redundancy between consecutive frames within the multi-frame stream of data; and the first frame includes the raw image.

[0173] According to an embodiment to the disclosure, to map the raw image to the compressed domain, the at least one processing device is configured to order data from most significant bits to least significant bits, or to perform the one or more image signal processing operations, the at least one processing device is configured to process only the most significant bits of data in the compressed domain to perform at least one of the one or more image signal processing operation that does not require full-resolution data.

[0174] Although this disclosure has been described with reference to various example embodiments, various changes and modifications may be suggested to one skilled in the art. It is intended that this disclosure encompass such changes and modifications as fall within the scope of the appended claims.

Claims

1.A method for image signal processing, the method being performed by an electronic device (101) and comprising:obtaining (201), by at least one processor (120) of the electronic device (101), a raw image (302);mapping (202), by the at least one processor (120), using a raw image encoder (308), the raw image (302) to a compressed domain, wherein the raw image (302) is represented using latent variables in the compressed domain;performing (203), by the at least one processor (120), one or more image signal processing operations on the latent variables, wherein each of the one or more image signal processing operations is configured to operate in the compressed domain, and wherein the one or more image signal processing operations generate processed latent variables; andmapping (204), by the at least one processor (120), using an output image decoder (305), the processed latent variables to an output image (306) in an output color space.2.The method of claim 1, wherein the output color space is a red-green-blue (RGB) color space and the output image (306) is an RGB image; orthe output color space is a luminance-chrominance (YUV) color space and the output image (306) is a YUV image.3.The method of claim 1 or 2, wherein the raw image encoder (308) comprises a convolutional neural network-based encoder with quantization; orthe output image decoder (305) comprises a convolutional neural network.4.The method of any one of claims 1 to 3, wherein the one or more image signal processing operations comprise at least one of: compressed domain denoising, compressed domain sharpening, compressed domain point spread function (PSF) inversion, compressed domain segmentation, compressed domain motion map estimation and motion compensation, or compressed domain motion image registration.5.The method of any one of claims 1 to 4, wherein the raw image encoder (308) comprises a trained machine learning model subjected to a training scheme comprising:training a teacher image encoder (303) to map input images into the compressed domain, wherein the training yields trained network weights for the teacher image encoder (303);initializing network weights of the raw image encoder (308) from the trained network weights for the teacher image encoder (303); andconstraining outputs of the raw image encoder (303) to be close to outputs of the teacher image encoder (303) via a mean squared error loss in the compressed domain.6.The method of any one of claims 1 to 5, wherein:the teacher image encoder comprises a red-green-blue (RGB) image encoder; andthe training scheme comprises training the teacher image encoder to map RGB images into the compressed domain.7.The method of any one of claims 1 to 6, wherein the mapping the raw image to the compressed domain comprises:mapping a multi-frame stream of data including the raw image to the compressed domain based on (i) a first frame within the multi-frame stream of data and (ii) at least one of:spatial redundancy between the first frame and a remainder of frames within the multi-frame stream of data; orthe spatial redundancy between the first frame and the remainder of frames and temporal redundancy between consecutive frames within the multi-frame stream of data; andwherein the first frame includes the raw image.8.An electronic device (101) for image signal processing, the electronic device comprising:at least one processing device (120) configured to:obtain a raw image (302);map, using a raw image encoder (308), the raw image (302) to a compressed domain, wherein the raw image (302) is represented using latent variables in the compressed domain;perform one or more image signal processing operations on the latent variables, wherein each of the one or more image signal processing operations is configured to operate in the compressed domain, and wherein the one or more image signal processing operations are configured to generate processed latent variables; andmap, using an output image decoder (305), the processed latent variables to an output image (306) in an output color space.9.The electronic device of claim 8, wherein one of:the output color space is a red-green-blue (RGB) color space and the output image (306) is an RGB image; orthe output color space is a luminance-chrominance (YUV) color space and the output image (306) is a YUV image.10.The electronic device of any one of claim 8 or 9, wherein at least one of:the raw image encoder comprises a convolutional neural network-based encoder with quantization; orthe output image decoder comprises a convolutional neural network.11.The electronic device of any one of claims 8 to 10, wherein the one or more image signal processing operations comprise at least one of: compressed domain denoising, compressed domain sharpening, compressed domain point spread function (PSF) inversion, compressed domain segmentation, compressed domain motion map estimation and motion compensation, or compressed domain motion image registration.12.The electronic device of any one of claims 8 to 11, wherein the raw image encoder comprises a trained machine learning model subjected to a training scheme comprising:training a teacher image encoder to map input images into the compressed domain, wherein the training yields trained network weights for the teacher image encoder;initializing network weights of the raw image encoder from the trained network weights for the teacher image encoder; andconstraining outputs of the raw image encoder to be close to outputs of the teacher image encoder via a mean squared error loss in the compressed domain.13.The electronic device of any one of claims 8 to 12, wherein:to map the raw image to the compressed domain, the at least one processing device is configured to map a multi-frame stream of data including the raw image to the compressed domain based on (i) a first frame within the multi-frame stream of data and (ii) at least one of:spatial redundancy between the first frame and a remainder of frames within the multi-frame stream of data; orthe spatial redundancy between the first frame and the remainder of frames and temporal redundancy between consecutive frames within the multi-frame stream of data; andthe first frame includes the raw image.14.The electronic device of any one of claims 8 to 13, whereinto map the raw image to the compressed domain, the at least one processing device is configured to order data from most significant bits to least significant bits; orto perform the one or more image signal processing operations, the at least one processing device is configured to process only the most significant bits of data in the compressed domain to perform at least one of the one or more image signal processing operations that does not require full-resolution data.15.A computer-readable storage medium storing instructions, wherein, the instructions, when executed by at least one processor, cause the at least one processor to perform the method according to any one of claims 1 to 7.

Images