Image quality processing method and apparatus, electronic device, and storage medium
By processing and pre-coding the initial image quality on the cloud, the image quality loss caused by cloud rendering is solved, and the image quality and user experience of cloud rendering images are improved.
Patent Information
- Application Number
- PCT/CN2024/089664
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-06-01
- Filing Date
- 2024-04-24
- Publication Date
- 2024-12-05
AI Technical Summary
In application scenarios such as cloud gaming or metaverse, the image quality loss caused by cloud rendering makes the user's terminal device's picture quality experience less than the local rendering effect. The existing technology only improves the encoder parameters by adjusting only.
By obtaining the initial image texture data, converting it into target image texture data suitable for the processor computing architecture, the image quality enhancement process is performed using the image quality processing algorithm in the processor computing architecture, and the image is color enhanced before encoding and Denoising processing, adjust the encoder parameters to allocate bandwidth according to the area of interest.
It significantly improves the image quality of cloud rendered images, improves user experience, and achieves higher post-encoded image quality.
Smart Images

Figure CN2024089664_05122024_PF_FP_ABST
Abstract
Description
Image quality processing method and device, electronic device and storage medium
[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on June 1, 2023, with application number 202310647859.0 and invention name “A method and device for image quality processing, electronic device and storage medium”, the entire contents of which are incorporated by reference into this application. Technical Field
[0002] The present application relates to the field of cloud computing technology, and in particular to a method and device for image quality processing, an electronic device, and a storage medium. Background Art
[0003] In cloud gaming or metaverse applications, cloud rendering is often used to render images such as cloud gaming pages or metaverse pages. Cloud rendering involves rendering on a cloud server. Control commands are sent from the user terminal to the cloud, and the cloud server performs the corresponding rendering tasks based on the control commands. The rendered image is compressed and encoded before being sent back to the user terminal for display.
[0004] However, the compression encoding process often results in a loss of image quality, making the image quality experience on the user's terminal device inferior to that of local rendering. Related technologies rely solely on adjusting encoder parameters to improve image quality, but the improvement is limited.
[0005] Therefore, there is an urgent need to provide a method to improve image quality and enhance the user experience of cloud rendering.
[0006] Summary of the Invention
[0007] The embodiments of the present application provide a method and device for image quality processing, an electronic device, and a storage medium to solve one or more of the above-mentioned technical problems.
[0008] In a first aspect, an embodiment of the present application provides a method for image quality processing, including: obtaining initial image texture data; converting the initial image texture data into target image texture data suitable for a processor computing architecture; performing image quality enhancement processing on the target image texture data using an image quality processing algorithm in the processor computing architecture; and sending the target image texture data after image quality enhancement processing to an encoder, so that the encoder encodes the image based on the target image texture data after image quality enhancement processing.
[0009] In the second aspect, an embodiment of the present application provides an image processing method implemented by interacting with a cloud server, including: obtaining a picture acquisition component provided by the cloud server, and injecting the picture acquisition component into the image rendering engine of the image loading program; capturing the image rendered by the image rendering engine through the hook function of the picture acquisition component, and providing the initial image texture data corresponding to the image to the cloud server for the cloud server to perform image encoding; wherein, the cloud server converts the initial image texture data into target image texture data suitable for the processor computing architecture; the target image texture data uses the image quality processing algorithm of the processor computing architecture to perform image quality enhancement processing; the target image texture data after image quality enhancement processing is sent to the encoder, so that the encoder encodes the image based on the target image texture data after image quality enhancement processing.
[0010] In a third aspect, an embodiment of the present application provides an image quality processing device, including: a texture data acquisition module for acquiring initial image texture data; a texture data conversion module for converting the initial image texture data into target image texture data suitable for a processor computing architecture; an image quality enhancement processing module for performing image quality enhancement processing on the target image texture data using an image quality processing algorithm in the processor computing architecture; and a texture data sending module for sending the target image texture data after image quality enhancement processing to an encoder, so that the encoder performs encoding processing on the image based on the target image texture data after image quality enhancement processing.
[0011] In a fourth aspect, an embodiment of the present application provides an image processing device implemented by interacting with a cloud server, comprising: an acquisition component injection module for acquiring a picture acquisition component provided by the cloud server, and injecting the picture acquisition component into the image rendering engine of the image loading program; a texture data providing module for capturing the image rendered by the image rendering engine through the hook function of the picture acquisition component, and providing the initial image texture data corresponding to the image to the cloud server for the cloud server to perform image encoding; wherein, the cloud server converts the initial image texture data into target image texture data suitable for the processor computing architecture; the target image texture data is subjected to picture quality enhancement processing using the picture quality processing algorithm of the processor computing architecture; the target image texture data after picture quality enhancement processing is sent to the encoder, so that the encoder encodes the image based on the target image texture data after picture quality enhancement processing.
[0012] In a fifth aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor implements any of the above-described methods when executing the computer program.
[0013] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements any of the methods described above.
[0014] The above description is only an overview of the technical solution of this application. In order to more clearly understand the technical means of this application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of this application more obvious and easy to understand, the specific implementation methods of this application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the multiple drawings represent the same or similar components or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments according to the present application and should not be regarded as limiting the scope of the present application.
[0016] FIG1 shows a flow chart of an application example of an image quality processing solution provided in an embodiment of the present application.
[0017] FIG2 shows a flow chart of an image quality processing method provided in an embodiment of the present application.
[0018] FIG3 shows a flowchart of an image processing method implemented by interacting with a cloud server, provided in an embodiment of the present application.
[0019] FIG4 shows a structural block diagram of an image quality processing device provided in an embodiment of the present application.
[0020] FIG5 shows a structural block diagram of an image processing device implemented by interacting with a cloud server, provided in an embodiment of the present application.
[0021] FIG6 shows a block diagram of an electronic device for implementing an embodiment of the present application. DETAILED DESCRIPTION
[0022] Hereinafter, only certain exemplary embodiments are briefly described. As will be appreciated by those skilled in the art, the described embodiments may be modified in various ways without departing from the spirit or scope of the present application. Therefore, the drawings and description are to be regarded as illustrative in nature and not restrictive.
[0023] To facilitate understanding of the technical solutions of the embodiments of the present application, the following describes the related technologies of the embodiments of the present application. The following related technologies can be combined with the technical solutions of the embodiments of the present application as optional solutions, and all of them fall within the scope of protection of the embodiments of the present application.
[0024] In a related technology prior to this application, image quality was improved solely by adjusting encoder parameters, such as the encoder bit rate, and using the same quantization parameter (QP value), which evenly distributed bandwidth across all image regions. This approach cannot use image quality processing algorithms to perform pre-encoding processing such as color enhancement and denoising on the original rendered image before encoding to improve image quality, nor does it allocate bandwidth to different image regions based on points of interest. This results in limited improvement in image quality and a poor user experience.
[0025] In view of this, embodiments of the present application provide an image quality processing method and device, an electronic device, and a storage medium to solve all or part of the above-mentioned technical problems.
[0026] Figure 1 is a flow chart illustrating an exemplary application example for implementing an image quality processing solution provided in an embodiment of the present application. When cloud rendering is performed on real-time images generated in a cloud-based online game or metaverse, the cloud rendering device deployed in the cloud can include three parts: an image rendering engine (render.exe), a GPU graphics processing unit, and a cloud streaming component (Streamer.exe). These components can all be deployed on the servers of vendors providing cloud computing services. The image rendering engine can use its texture rendering component deployed in the cloud to render the image to be rendered, and send the rendered image data to the GPU graphics processing unit in the cloud; after receiving the rendered image data, the graphics processing unit can store it in the background buffer of the graphics processing unit memory exchange chain, and can further render the image data stored in the background buffer into image texture data, and encode the image texture data through a hardware encoder, and then send the encoded image to the encoder in the cloud streaming component (the encoder in the cloud streaming component can be a logical abstraction of the hardware encoder); in this process, the cloud streaming component can call the function interface of the corresponding module in the GPU image processing unit, use the computing power of the GPU graphics processing unit to process the image texture data, and after receiving the encoded image, send the image to the user terminal device via network transmission.
[0027] In order to improve the image quality before encoding, a cloud streaming component can be used to remotely inject a picture acquisition component into the texture rendering component of the image rendering engine before encoding the image. The picture acquisition component can include a hook function, which can capture the image output by the image rendering function rendered by the image rendering engine and extract the initial image texture data corresponding to the image. After capturing the rendered image and extracting the initial image texture data, the cloud streaming component can call a function interface for data conversion between the texture rendering component (DirectX) and the processor computing architecture (which can be a general parallel computing architecture CUDA) to convert the initial image texture data into target image texture data suitable for the processor computing architecture (data conversion), so that the format of the target image texture data after data conversion can conform to the data format required for image quality enhancement processing using an image quality processing algorithm, such as a deep learning inference engine tensorRT, and / or an image quality enhancement algorithm based on traditional vision, such as a CAS algorithm (contrast adaptive sharpening, contrast adaptive sharpening algorithm), etc., and then the cloud streaming component can call the image quality processing algorithm function interface to use the image quality processing algorithm in the processor computing architecture to perform image quality enhancement processing on the target image texture data. Therefore, the cloud streaming component can connect the image rendering engine, GPU graphics processing unit and user terminal in series, and can use the image quality processing algorithm to pre-encode the image before encoding and sending it to the user terminal device. For example, the image quality processing algorithm can be used to perform color enhancement and / or denoising on the image before encoding, thereby significantly improving the image quality of the encoded image and enhancing the user experience.
[0028] The execution subject in the embodiment of the present application may be the aforementioned cloud streaming component (Streamer.exe). As shown in Figure 1, the cloud streaming group may be composed of three parts, namely video capture, network security mechanism (QoS), and encoder. Among them, video capture can be used to capture rendered images, which may include a batch rendering management plug-in (render manager) and a screen capture component (DLL injector); the network security mechanism can be used to solve problems such as network delay and congestion, and ensure the real-time performance of cloud rendering; the encoder may be a logical abstraction of a hardware encoder, and the hardware encoder may compress and encode the rendered image so that it can adapt to the user terminal device. The cloud streaming component may transmit the encoded image to the user terminal via the network.
[0029] First, you can use the cloud streaming component to obtain the initial image texture data rendered by the image rendering engine of the image loader (which can be a game program). The image rendering engine uses the texture rendering component to render the image to be generated. In the texture rendering component, you can deploy rendering units with different functions, such as the input assembly unit (Input Assembler, which can be used to parse the input data), the vertex shader (Vertex Shader, which can be used to add various special effects), the hull shader (which can generate the surface control points and tessellation parameters of each primitive and its topological structure based on factors such as vectors), the blocking unit (Tessellator, which can divide some large and disordered graphics into several small graphics, and then combine the small graphics to form ordered geometric graphics, improve the quality, performance, and controllability of the graphics, and make the graphics closer to reality. Tessellation Shader is also called tessellation shader), domain shader (Domain Shader, which can process separated points into graphics), geometry shader (Geometry Shader, which can call a single primitive as input and can output multiple primitives or no primitives). After being processed by the above-mentioned shell shader, blocking unit, domain shader, and geometry shader, the output value of the vertex shading engine can be clipped, perspective-split, and viewport transformed, thus further entering the rasterization stage. The rendering units with different functions deployed in the texture rendering component can also include a rasterizer (which can rasterize primitives), a pixel shader (which can process the pixels inside the triangle after rasterization), and an output merge unit (which can make depth judgments on pixel displays).
[0030] When the image rendering engine uses the texture rendering component to render the image to be generated, the above-mentioned rendering unit can be used to render the image (see the process corresponding to the arrow marked ② in Figure 1), and the rendered image data is sent to the GPU graphics processing unit in the cloud. The way to send image data is that after the image rendering engine renders a frame of image, the image rendering engine calls the image rendering function (present function) (see the process corresponding to the arrow marked ③ in Figure 1), and the image rendering function outputs the rendered image to the GPU graphics processing unit, which then stores the rendered image in the background buffer in the graphics processing unit memory swap chain. Therefore, this feature of the image rendering engine can be used to remotely inject a screen capture component (DLL injector) containing a hook function into the image rendering function (see the process corresponding to the arrow marked ① in Figure 1), wherein the hook function can be used to capture the image output by the image rendering function (present function) to the GPU graphics processing unit. It can be that when the image rendering engine calls the Present function, the hook function in the screen capture component of the cloud streaming component captures the image rendered when the Present function is executed, and then extracts the initial image texture data corresponding to the image (which can be DirectX texture data). When remotely injecting a screen capture component containing a hook function into the image rendering function, an open source function library can be used for injection, such as the detours library (MicroSoft Detours, a function library for capturing system APIs), the EasyHook library (a hook function library, a function library for capturing system APIs, generally having three hierarchical structures: a HOOK API assembly layer, a HOOK management layer, and an application layer that replaces API functions), etc.
[0031] After the GPU graphics processing unit stores the image rendered by the image rendering engine in the back buffer of the GPU memory swap chain, it can render the initial image texture data (rendered texture, shared texture) corresponding to the image in the back buffer in the GPU memory (VRAM) (see the process corresponding to the arrow labeled ④ in Figure 1), and then send the initial image texture data out through the GPU graphics processing unit's hardware encoder. In the GPU memory swap chain, there can be multiple back buffers, such as back buffer 1, back buffer 2, and back buffer 3 in Figure 1. The image rendered by the image rendering engine can be first stored in back buffer 1, and the GPU graphics processing unit can first render the texture data corresponding to the image stored in back buffer 2. After rendering the texture data corresponding to the image stored in back buffer 2, the image rendered by the image rendering engine can be stored in back buffer 3, while simultaneously rendering the texture data corresponding to the image stored in back buffer 1. After rendering the image in back buffer 1, the image rendered by the image rendering engine can be stored in back buffer 2, while simultaneously rendering the texture data corresponding to the image stored in back buffer 3. Similarly, different background buffers can be used alternately to store the images rendered by the game calculations and the texture data corresponding to the rendered images, thereby ensuring the smoothness of the game images when displayed on the user terminal.
[0032] Secondly, after using the cloud streaming component to obtain the initial image texture data rendered by the image rendering engine of the image loader, the function interface for data conversion between the texture rendering component of the image rendering engine and the processor computing architecture (data conversion function interface, see the process corresponding to the arrow labeled ⑤ in Figure 1) can be called to convert the initial image texture data into target image texture data suitable for the processor computing architecture. For example, when calling the data conversion function interface, the GPU graphics processing unit performs data conversion on the initial image texture data (which can be DirectX texture data) obtained by the cloud streaming component (see the process corresponding to the arrow labeled ⑥ in Figure 1) and converts it into target image texture data (which can be CUDA data). That is, the DirectX texture can be mapped to CUDA data, and then the target image texture data can be placed on the general parallel computing architecture (CUDA) for image quality enhancement processing.
[0033] Again, the image quality processing algorithm function interface can be called through the cloud streaming component, and the image quality processing algorithm in the processor computing architecture can be used to perform image quality enhancement processing on the target image texture data. The format of the target image texture data after the previous step conversion conforms to the data format required for image quality enhancement processing using the image quality processing algorithm on cuda. Among them, the image quality processing algorithm can adopt a deep learning inference engine (deep learning tensorRT), and / or, a traditional vision-based image quality enhancement algorithm such as CAS algorithm (contrast adaptive sharpening, contrast adaptive sharpening algorithm) and the like. For example, image quality enhancement processing on a general parallel computing architecture (cuda) may have restrictions on the texture data format of the image, such as not supporting formats such as R10G10B10A2, etc. Therefore, RTT (Render To Texture) technology can be used to convert the format into a format supported by the general parallel computing architecture (cuda), such as B8G8R8A8 format, etc. Furthermore, the image texture data in the converted format can be enhanced using the deep learning inference engine tensorRT and / or traditional vision-based image quality enhancement algorithms such as the CAS algorithm (contrast adaptive sharpening) on a general-purpose parallel computing architecture (cuda). Otherwise, the above tools cannot be used for image quality enhancement.
[0034] Finally, the target image texture data after image quality enhancement processing can be sent to the encoder, so that the encoder can encode the image based on the target image texture data after image quality enhancement processing. In addition, an image quality processing algorithm or other algorithm can be used on the processor computing architecture to assign corresponding quantization parameters to the image corresponding to the target image texture data, and the quantization parameters and the target image texture data can be sent to the encoder together, so that the encoder can configure the encoding parameters according to the quantization parameters. For example, the image can be identified as an object, and the area of interest in the image can be determined. According to the area of interest in the image, corresponding quantization parameters can be assigned to different areas of the image. That is, a reasonable QP distribution (bandwidth distribution of different areas of the image) for each frame of the image can be calculated, and the QP distribution can be output to the encoder to adjust the encoder parameters so that the encoder encodes the image according to the parameters.
[0035] Therefore, an embodiment of the present application provides an image quality processing solution that includes two types of processing logic: pre-encoding processing and encoding parameter control. The pre-encoding processing logic is to convert the initial image texture data rendered by the image rendering engine into a texture data format that can be processed using an image quality processing algorithm. The image quality processing algorithm is used to improve the image quality before encoding, and the cloud-rendered image is processed by algorithms such as color enhancement and / or denoising. The encoding parameter control processing logic is to adjust the encoder's encoding parameters according to the image area, adjusting the compression rate and bandwidth of different areas in the image. Using these two types of processing logic to process the image quality of the cloud-rendered image simultaneously can improve the image quality of the cloud rendering and enhance the end user experience.
[0036] The execution subject of the embodiment of the present application can be an application, service, instance, software-based functional module, virtual machine (VM), container or cloud server, etc., or a hardware device with data processing function (such as a server or terminal device) or hardware chip (such as CPU, GPU, FPGA, NPU, AI acceleration card or DPU), etc. The device for implementing image quality processing can be deployed on the computing device of the application party that provides the corresponding service or on a cloud computing platform that provides computing power, storage and network resources. The mode in which the cloud computing platform provides services to the outside world can be IaaS (Infrastructure as a Service), PaaS (Platform as a Service), SaaS (Software as a Service) or DaaS (Data as a Service). Taking the platform providing SaaS software as a service (Software as a Service) as an example, the cloud computing platform can use its own computing resources to provide training of the image quality processing model or functional execution of the image quality processing module, and the specific application architecture can be built according to service requirements. For example, the platform can provide construction services based on the above model to applications or individuals using platform resources, and further call the above model and implement online or offline image quality processing functions based on image quality processing requests submitted by relevant clients or servers and other devices.
[0037] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.
[0038] The following describes in detail the technical solution of this application and how it solves the aforementioned technical problems using specific embodiments. The several specific embodiments listed can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. The following describes the embodiments of this application in detail with reference to the accompanying drawings.
[0039] An embodiment of the present application provides an image quality processing method. FIG2 is a flowchart of an image quality processing method 200 according to an embodiment of the present application. The method 200 may include the following steps S201 to S204 .
[0040] In step S201 , initial image texture data is obtained.
[0041] The initial image texture data involved in the embodiments of the present application may be texture data corresponding to a rendered game image, and may be used to characterize image-related features. Texture may be a visual feature that reflects homogeneous phenomena in an image, and it reflects the slowly changing or periodically changing surface structural organization and arrangement properties of an object's surface, and may be represented by the grayscale distribution of pixels and their surrounding spatial neighborhoods. While reflecting the properties of global features, texture features may also describe the surface properties of the scene corresponding to an image or image region, and may be statistically calculated in an area containing multiple pixels.
[0042] In a possible implementation, the initial image texture data may be obtained by obtaining an image captured by a hook function of a screen acquisition component in an image rendering engine of an image loading program, and extracting the initial image texture data corresponding to the image.
[0043] The image loading program involved in the embodiments of the present application may be a loading program that uses cloud computing resources to perform calculations and generate images, such as a game program or a metaverse program. Such programs can be deployed on a cloud server and render images through cloud rendering. The rendered image can be an image frame in a video, or it can be an image generated according to instructions from a user terminal by interacting with the user terminal. The image rendering engine involved can be deployed on a server of a manufacturer that provides cloud computing services, and is used to render corresponding game images, etc. The image rendering engine can use the render.exe process for rendering.
[0044] For example, when a user plays a cloud game on a mobile phone, the cloud game program (i.e., the image loading program) is deployed in a cloud server. If the user clicks the forward button on the mobile phone (the purpose may be to make the corresponding character in the cloud game move forward), the mobile server will send the instruction of the user clicking the forward button to the cloud server through the network. After the cloud server receives the instruction, it can send the instruction to the image loading program. The image loading program responds to the instruction and makes the corresponding character in the cloud game move forward. At this time, a new game screen will be formed. Therefore, the image rendering engine of the image loading program will render the game screen. At the same time, after receiving the aforementioned forward instruction, the cloud server can also send the instruction to the cloud streaming component through the agent program in the server, thereby starting the cloud streaming component. After the image rendering engine renders the game screen, the cloud streaming component can obtain the image rendered by the image rendering engine of the image loading program, and then use the encoder to encode it (for example, into a 264 or a 265 video code stream), and transmit the rendered image back to the user's terminal mobile phone.
[0045] Among them, the picture acquisition component can be injected into the image rendering engine in advance through the cloud streaming component before the rendered image is encoded, and is used to capture the image rendered in the image rendering engine. A DLL injector can be used (DLL injector is used to force program A to load a.dll given by program B, so that a.dll enters the address space of program A, and then program A executes the code in a.dll given by program B). In the picture acquisition component, a hook function can be included. When the picture acquisition component is injected into the image rendering engine, the hook function in the picture acquisition component can be registered with the corresponding image rendering function of the image rendering engine, such as the present function. At this time, if the image rendering engine calls the image rendering function (present function) when rendering the image frame, the rendered image can be captured. When the image rendering engine renders the image frame, each time a frame of image is rendered, the image rendering function (present function) will be called to pass the rendered image frame to the GPU graphics processing unit. Therefore, this feature of the image rendering engine can be used to capture the image.
[0046] In some embodiments, before obtaining the initial image texture data, a screen acquisition component can be remotely injected into the image rendering engine. The screen acquisition component can include a hook function, which can be used to capture the image output by the image rendering function rendered by the image rendering engine.
[0047] Accordingly, the method of injecting a screen capture component (DLL injector) containing a hook function into the image rendering engine can be a remote injection method (a program injects into another program other than itself), that is, injecting the image rendering engine from another cloud program, such as injecting from a cloud serial component into the image rendering engine. The hook function can be remotely injected into the image rendering function of the image rendering engine. Among them, the hook function can be used to capture the image output by the image rendering function (present function) to the GPU graphics processing unit. When the image rendering engine calls the Present function, the hook function in the screen capture component of the cloud streaming component can capture the image rendered when the Present function is executed, and then extract the initial image texture data corresponding to the image (which can be DirectX texture data). When remotely injecting a screen acquisition component containing a hook function into an image rendering function, you can use an open source function library for injection, such as the detours library (MicroSoft Detours, a function library for capturing system APIs), the EasyHook library (a hook function library, a function library for capturing system APIs, generally has three hierarchical structures: the HOOK API assembly layer, the HOOK management layer, and the application layer that replaces the API function), etc.
[0048] In step S202 , the initial image texture data is converted into target image texture data suitable for the processor computing architecture.
[0049] Due to data format limitations, the initial image texture data must be converted into target image texture data suitable for the processor computing architecture before the image quality processing algorithm in the processor computing architecture can be used to enhance the image quality of the target image texture data. Therefore, data conversion of the initial image texture data is required.
[0050] The processor computing architecture involved is a hardware and software architecture for operating processors such as CPUs and GPUs. It provides a simple interface for programming processors and building processor-based applications. The processor computing architecture can be used to perform various processor-based calculations and deploy various algorithms that call processor execution on the processor computing architecture, thereby leveraging the processing power of the processor to significantly improve computing performance.
[0051] In one possible implementation, the initial image texture data is converted into target image texture data suitable for a processor computing architecture. A function interface for performing data conversion between a texture rendering component of an image rendering engine and the processor computing architecture can be called to convert the initial image texture data into target image texture data suitable for the processor computing architecture, wherein the processor computing architecture may include a general parallel computing architecture.
[0052] In the embodiments of the present application, the general parallel computing architecture (CUDA) involved can be a software and hardware architecture for operating GPU computing, which can perform various calculations of the GPU. The software components of CUDA are: a CUDA library, an application programming interface and its runtime library, and two higher-level general-purpose math libraries. CUDA improves read and write flexibility and provides on-chip shared memory so that data can be shared between threads. Applications can use shared memory to reduce data transfer and rely less on memory bandwidth. On the general parallel computing architecture (CUDA), image quality processing algorithms that call GPU execution can be deployed, such as the deep learning inference engine tensorRT and / or image quality enhancement algorithms based on traditional vision such as CAS algorithm (contrast adaptive sharpening, contrast adaptive sharpening algorithm) and the like.
[0053] The texture rendering component involved may be DirectX, such as DirectX12, DirectX11, etc. This component can store 2D image data. In the image texture data in the DirectX format, each element can store the color of a pixel, or it can be a 3D vector. For example, the format DXGI FORMAT R32G32B32 FLOAT contains 3 floating-point components, so any 3D vector can be stored in a floating-point coordinate format. Image quality enhancement processing on a general parallel computing architecture (cuda) may have restrictions on the texture data format of the image, such as not supporting formats such as R10G10B10A2, etc. Therefore, the RTT (Render To Texture) technology can be used to convert the format into a format supported by the general parallel computing architecture (cuda), such as the B8G8R8A8 format. Furthermore, the image texture data in the converted format can be enhanced using the deep learning inference engine tensorRT and / or traditional vision-based image quality enhancement algorithms such as the CAS algorithm (contrast adaptive sharpening) on a general-purpose parallel computing architecture (cuda). Otherwise, the above tools cannot be used for image quality enhancement.
[0054] The function interface called for data conversion between the texture rendering component of the image rendering engine and the processor computing architecture may be cudaGraphicsD3D11RegisterResource, a function interface for converting DirectX image texture data to CUDA resources. The image texture data in DirectX format before conversion using the above function interface may be referred to as "initial image texture data," and the image texture data in CUDA format after conversion may be referred to as "target image texture data."
[0055] In step S203 , the image quality enhancement process is performed on the target image texture data using the image quality processing algorithm in the processor computing architecture.
[0056] In one possible implementation, the image quality enhancement processing of the target image texture data using the image quality processing algorithm of the processor computing architecture can be performed by using a deep learning inference engine pre-installed in the processor computing architecture and / or an image quality enhancement algorithm based on traditional vision such as a CAS algorithm (contrast adaptive sharpening, contrast adaptive sharpening algorithm) to enhance the image quality of the target image texture data.
[0057] In an embodiment of the present application, the deep learning inference engine pre-installed in the processor computing architecture (such as the general parallel computing architecture cuda) can be a tensorRT model, which can enhance the image quality of the target image texture data by combines layers (combined layers), kernel (kernel) optimization selection, etc., and can also perform normalization and conversion into a better matrix math (matrix editing) method for image quality enhancement according to the specified accuracy. The model can also be trained in advance to enable developers to import, calibrate, generate and deploy optimized networks so that the tensorRT model can perform optimization operations (i.e., image quality enhancement) according to the definition of the network. Of course, the deep learning inference engine pre-installed in the processor computing architecture can also be other deep learning models, and this application does not impose any restrictions on this.
[0058] The image quality enhancement algorithm based on traditional vision that is pre-installed in the processor computing architecture can enhance the image quality using this type of algorithm. For example, the brightness, contrast, saturation, hue, etc. of the image can be adjusted to increase the clarity of the image, enhance the color of the image, reduce the noise of the image, etc. For example, a CAS algorithm (contrast adaptive sharpening algorithm) can be used. This algorithm can adjust the degree of sharpening according to the contrast of the image. For example, the algorithm can adjust the sharpening amount of each pixel to sharpen the clearer areas of the image less and the areas lacking details more, so that the entire image achieves uniform sharpness, reduces artifacts in the image, improves the clarity of the image, and thus improves the overall visual effect of the game screen or the metaverse screen, so that the end user can feel more details in the image. Of course, the image quality enhancement algorithm based on traditional vision that is pre-installed in the processor computing architecture can also adopt other algorithms, and this application does not impose any restrictions on this.
[0059] In addition, other algorithms for enhancing image texture can be preset in the processor computing architecture, and then these preset algorithms can be used to enhance the image quality of the target image texture data. For example, the gray level co-occurrence matrix (GLCM) can be used to enhance texture features based on the relationship between different pixels in the image; the Gabor filter can be used to extract texture information of different scales and directions for enhancement; the local binary pattern (LBP) can be used to calculate the difference between a pixel and its neighboring pixels, and the texture information can be enhanced by encoding it into a binary number; the histogram of directional gradients (HOG) can be used to calculate the gradient information of different directions in the image, and encode it into a histogram to further enhance the texture information, etc. This application does not impose any restrictions on this.
[0060] Accordingly, the aforementioned image quality enhancement processing may include color enhancement and / or denoising processing. The color enhancement and / or denoising processing of an image may be performed using a single image quality processing algorithm, or may be performed using multiple image quality processing algorithms simultaneously or sequentially, and this application does not impose any restrictions on this.
[0061] In addition to color enhancement and / or denoising, you can also use Gaussian filtering to smooth the image and enhance the texture. The degree of smoothing and enhancement can be controlled by adjusting the filter size and standard deviation. Median filtering can be used to remove noise and enhance the texture of the image, replacing the value of each pixel with the median of its surrounding pixels, preserving edges and details. Sobel filtering can be used to detect edges in the image and enhance the texture. It can be applied in both horizontal and vertical directions to detect horizontal and vertical edges in the image. Laplacian filtering can be used to detect edges and details in the image, highlighting them to enhance the texture. Morphological operations (dilation and erosion) can be used to enhance the texture and details in the image. This application does not impose any restrictions on this.
[0062] In step S204 , the target image texture data after the image quality enhancement processing is sent to the encoder, so that the encoder performs encoding processing on the image based on the target image texture data after the image quality enhancement processing.
[0063] In one possible implementation, the target image texture data after image quality enhancement processing is sent to the encoder by first assigning a corresponding quantization parameter to the image corresponding to the target image texture data, and then sending the quantization parameter together with the target image texture data to the encoder, so that the encoder can configure the encoder's encoding parameters according to the quantization parameter.
[0064] The quantization parameter assigned to the image corresponding to the target image texture data may be a QP distribution value. The QP value indicates the compression of the image's spatial details and, to a certain extent, determines the image quality. For example, a smaller QP value preserves more image details; a larger QP value may lose more image details, resulting in lower image clarity.
[0065] For example, the corresponding quantization parameters assigned to the image corresponding to the target image texture data can also be bandwidth parameters, pixel parameters, color parameters, light effect parameters, image compression rate and other parameters corresponding to different areas of the image. This application does not impose any restrictions on this.
[0066] In some embodiments, the above-mentioned image corresponding to the target image texture data is assigned corresponding quantization parameters. The image can first be subjected to object recognition to determine the interest point area in the image, and then the corresponding quantization parameters are assigned to different areas in the image according to the interest point area in the image.
[0067] Based on the points of interest in the image, different quantization parameters can be assigned to different areas of the image. For example, the points of interest in the image can be calculated, and based on the calculation results, more bandwidth can be allocated to image areas with high points of interest to the end user (such as a cloud game player), while less bandwidth can be allocated to image areas with low points of interest to the end user. Image areas allocated with more bandwidth will have less loss after compression encoding by the encoder; while image areas allocated with less bandwidth will have greater loss after compression encoding by the encoder.
[0068] In addition to the above adjustment methods, the encoder output bit rate can also be controlled by adjusting the discrete cosine transform quantization parameter size to output the target bit rate, so that the target image texture data is encoded by the encoder with less image loss.
[0069] An embodiment of the present application also provides an image processing method implemented by interacting with a cloud server. FIG3 is a flowchart of an image processing method 300 implemented by interacting with a cloud server in an embodiment of the present application. The method 300 may include the following steps S301 to S302.
[0070] In step S301 , a picture acquisition component provided by a cloud server is obtained, and the picture acquisition component is injected into an image rendering engine of an image loading program.
[0071] The executing entity of the embodiment of the present application can be a server or game program that renders the initial image texture data of the image rendering engine of the image loading program (such as cloud games). It can obtain the screen acquisition component provided by the cloud server and inject the screen acquisition component into the image rendering engine.
[0072] In step S302, the image rendered by the image rendering engine is captured through a hook function of the image acquisition component, and the initial image texture data corresponding to the image is provided to the cloud server for the cloud server to perform image encoding. The cloud server in the embodiment of the present application can be a cloud server that uses the aforementioned image quality processing solution in the embodiment of the present application to provide image quality enhancement processing services.
[0073] In which, the cloud server converts the initial image texture data into target image texture data suitable for the processor computing architecture; the target image texture data uses the image quality processing algorithm of the processor computing architecture to perform image quality enhancement processing; the target image texture data after image quality enhancement processing is sent to the encoder, so that the encoder encodes the image based on the target image texture data after image quality enhancement processing.
[0074] In a possible implementation, the above-mentioned image acquisition component may include a hook function, and the hook function is used to capture the image output by the rendering function of the image rendering engine of the image loading program.
[0075] According to an embodiment of the present application, the initial image texture data is first obtained, and then the initial image texture data is converted into target image texture data suitable for the processor computing architecture, and then the image quality processing algorithm in the processor computing architecture is used to perform image quality enhancement processing on the target image texture data, and finally the target image texture data after the image quality enhancement processing is sent to the encoder, so that the encoder encodes the image based on the target image texture data after the image quality enhancement processing. Using the above scheme, the initial image texture data rendered by the image rendering engine can be converted into a texture data format that can be processed using the image quality processing algorithm, so as to improve the image quality using the image quality processing algorithm before encoding. Compared with the method in the related art that only relies on adjusting the encoding parameters to improve the image quality, the embodiment of the present application can significantly improve the image quality of the encoded image under the same encoding parameters.
[0076] Corresponding to the application scenario and method of the method provided in the embodiment of the present application, the embodiment of the present application also provides an image quality processing device. As shown in Figure 4, it is a structural block diagram of an image quality processing device 400 of an embodiment of the present application. The device 400 may include: a texture data acquisition module 401 for acquiring initial image texture data; a texture data conversion module 402 for converting the initial image texture data into target image texture data suitable for a processor computing architecture; an image quality enhancement processing module 403 for performing image quality enhancement processing on the target image texture data using an image quality processing algorithm in the processor computing architecture; and a texture data sending module 404 for sending the target image texture data after image quality enhancement processing to an encoder, so that the encoder performs encoding processing on the image based on the target image texture data after image quality enhancement processing.
[0077] In one possible implementation, the texture data acquisition module 401 may include: an image capture submodule for acquiring an image captured by a hook function of a screen acquisition component in an image rendering engine of an image loading program; and a texture data extraction submodule for extracting initial image texture data corresponding to the image.
[0078] In some embodiments, before obtaining the initial image texture data, the above-mentioned device may also include: an acquisition component injection module, which is used to remotely inject a picture acquisition component into the image rendering engine, and the picture acquisition component includes a hook function, and the hook function is used to capture the image output by the image rendering function rendered by the image rendering engine.
[0079] In one possible implementation, the texture data conversion module 402 may include: a function interface calling submodule, used to call a function interface for performing data conversion between a texture rendering component of an image rendering engine and a processor computing architecture, and convert the initial image texture data into target image texture data suitable for the processor computing architecture, wherein the processor computing architecture includes a general parallel computing architecture.
[0080] In one possible implementation, the above-mentioned image quality enhancement processing module 403 may include: an image quality enhancement processing sub-module, which is used to perform image quality enhancement processing on the target image texture data using a deep learning inference engine pre-installed in the processor computing architecture and / or an image quality enhancement algorithm based on traditional vision.
[0081] In one possible implementation, the above-mentioned texture data sending module 404 may include: a quantization parameter allocation submodule, used to allocate corresponding quantization parameters to the image corresponding to the target image texture data; a texture data sending submodule, used to send the quantization parameters together with the target image texture data to the encoder, so that the encoder can configure the encoding parameters of the encoder according to the quantization parameters.
[0082] In some embodiments, the above-mentioned quantization parameter allocation submodule may include: an interest point area determination unit, used to perform object recognition on the image and determine the interest point area in the image; a quantization parameter allocation unit, used to allocate corresponding quantization parameters to different areas in the image according to the interest point area in the image.
[0083] In a possible implementation, the image quality enhancement processing includes color enhancement and / or denoising processing.
[0084] Corresponding to the application scenario and method of the method provided in the embodiment of the present application, the embodiment of the present application also provides an image processing device implemented by interacting with a cloud server. As shown in Figure 5, it is a structural block diagram of an image processing device 500 implemented by interacting with a cloud server in an embodiment of the present application. The device 500 may include: an acquisition component injection module 501, which is used to obtain the picture acquisition component provided by the cloud server and inject the picture acquisition component into the image rendering engine of the image loading program; a texture data providing module 502, which is used to capture the image rendered by the image rendering engine through the hook function of the picture acquisition component, and provide the initial image texture data corresponding to the image to the cloud server for the cloud server to perform image encoding; wherein, the cloud server converts the initial image texture data into target image texture data suitable for the processor computing architecture; the target image texture data uses the image quality processing algorithm of the processor computing architecture to perform image quality enhancement processing; the target image texture data after image quality enhancement processing is sent to the encoder, so that the encoder encodes the image based on the target image texture data after image quality enhancement processing.
[0085] In a possible implementation, in the above-mentioned device, the picture acquisition component includes a hook function, and the hook function is used to capture the image output by the rendering function of the image rendering engine of the image loading program.
[0086] The functions of each module in each device in the embodiments of the present application can be found in the corresponding description in the above method, and have corresponding beneficial effects, which will not be repeated here.
[0087] Figure 6 is a block diagram of an electronic device used to implement embodiments of the present application. As shown in Figure 6 , the electronic device includes a memory 601 and a processor 602. Memory 601 stores a computer program executable on processor 602. Processor 602 executes the computer program to implement the methods described in the above embodiments. Memory 601 and processor 602 may be present in one or more locations.
[0088] The electronic device further includes a communication interface 603 for communicating with external devices and performing data exchange transmission.
[0089] If the memory 601, processor 602, and communication interface 603 are implemented independently, the memory 601, processor 602, and communication interface 603 can be interconnected via a bus and communicate with each other. This bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. This bus can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, FIG6 shows only one thick line, but this does not mean that there is only one bus or only one type of bus.
[0090] In a specific implementation, if the memory 601 , the processor 602 and the communication interface 603 are integrated on a chip, the memory 601 , the processor 602 and the communication interface 603 can communicate with each other through an internal interface.
[0091] An embodiment of the present application provides a computer-readable storage medium storing a computer program, which implements the method provided in the embodiment of the present application when the program is executed by a processor.
[0092] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the method provided in any embodiment of the present application.
[0093] An embodiment of the present application also provides a chip, which includes a processor for calling and executing instructions stored in the memory from the memory, so that a communication device equipped with the chip executes the method provided in the embodiment of the present application.
[0094] An embodiment of the present application also provides a chip, including: an input interface, an output interface, a processor and a memory. The input interface, the output interface, the processor and the memory are connected through an internal connection path. The processor is used to execute the code in the memory. When the code is executed, the processor is used to execute the method provided in the embodiment of the present application.
[0095] It should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. It is worth noting that the processor may be a processor that supports the Advanced RISC Machines (ARM) architecture.
[0096] In one embodiment, the memory may include read-only memory and random access memory. The memory may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available. For example, static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM) and direct RAM bus random access memory (DR RAM).
[0097] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.
[0098] In the description of this specification, the reference terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless they are mutually inconsistent.
[0099] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the description of this application, "plurality" means two or more, unless otherwise specifically defined.
[0100] Any process or method described in the flowchart or otherwise described herein can be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a specific logical function or process. The scope of the preferred embodiments of the present application includes other implementations in which the functions may be performed in a different order than shown or discussed, including performing the functions substantially simultaneously or in reverse order depending on the functions involved.
[0101] The logic and / or steps described in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with such instruction execution systems, apparatuses or devices.
[0102] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the above embodiment method can be completed by instructing the relevant hardware through a program, which can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.
[0103] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the aforementioned integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium. The storage medium may be a read-only memory, a magnetic disk, or an optical disk, etc.
[0104] The above is merely an exemplary embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various modifications or substitutions within the technical scope described in this application, and such modifications or substitutions should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.
Claims
1. A method for image quality processing, comprising: Get initial image texture data; Converting the initial image texture data into target image texture data suitable for a processor computing architecture; Performing image quality enhancement processing on the target image texture data using an image quality processing algorithm in the processor computing architecture; The target image texture data after the image quality enhancement processing is sent to the encoder, so that the encoder performs encoding processing on the image based on the target image texture data after the image quality enhancement processing.
2. The method according to claim 1, wherein: The obtaining of initial image texture data comprises: Get the image captured by the hook function of the screen acquisition component in the image rendering engine of the image loader; Initial image texture data corresponding to the image is extracted.
3. The method according to claim 2, wherein: Before obtaining the initial image texture data, the method further includes: A picture acquisition component is remotely injected into the image rendering engine, wherein the picture acquisition component includes a hook function, and the hook function is used to capture an image output by an image rendering function rendered by the image rendering engine.
4. The method according to claim 1, wherein: The converting of the initial image texture data into target image texture data suitable for the processor computing architecture comprises: A function interface for data conversion between a texture rendering component of an image rendering engine and a processor computing architecture is called to convert the initial image texture data into target image texture data suitable for the processor computing architecture, wherein the processor computing architecture includes a general parallel computing architecture.
5. The method according to claim 1, wherein: The image quality enhancement processing of the target image texture data using the image quality processing algorithm of the processor computing architecture includes: The target image texture data is subjected to image quality enhancement processing using a deep learning inference engine pre-installed in the processor computing architecture and / or an image quality enhancement algorithm based on traditional vision.
6. The method according to claim 1, wherein: The sending of the target image texture data after image quality enhancement processing to the encoder comprises: Assigning a corresponding quantization parameter to an image corresponding to the target image texture data; The quantization parameter is sent together with the target image texture data to the encoder, so that the encoder configures the encoding parameters of the encoder according to the quantization parameter.
7. The method according to claim 6, wherein: The assigning of a corresponding quantization parameter to an image corresponding to the target image texture data comprises: Performing object recognition on the image to determine an area of interest in the image; Corresponding quantization parameters are allocated to different regions in the image according to the interest point regions in the image.
8. The method according to claim 1, wherein: The image quality enhancement processing includes color enhancement and / or noise removal processing.
9. An image processing method implemented by interacting with a cloud server, comprising: Obtaining a picture acquisition component provided by a cloud server, and injecting the picture acquisition component into an image rendering engine of an image loading program; Capturing the image rendered by the image rendering engine through the hook function of the image acquisition component, and providing the initial image texture data corresponding to the image to the cloud server, so that the cloud server can perform image encoding; The cloud server converts the initial image texture data into target image texture data suitable for the processor computing architecture; the target image texture data is subjected to image quality enhancement processing using the image quality processing algorithm of the processor computing architecture; The target image texture data after the image quality enhancement processing is sent to the encoder, so that the encoder performs encoding processing on the image based on the target image texture data after the image quality enhancement processing.
10. The method according to claim 9, wherein: The picture acquisition component includes a hook function, and the hook function is used to capture the image output by the rendering function of the image rendering engine of the image loading program.
11. An image quality processing device, comprising: A texture data acquisition module, used to acquire initial image texture data; A texture data conversion module, used to convert the initial image texture data into target image texture data suitable for a processor computing architecture; An image quality enhancement processing module, used for performing image quality enhancement processing on the target image texture data using an image quality processing algorithm in the processor computing architecture; The texture data sending module is used to send the target image texture data after the image quality enhancement processing to the encoder, so that the encoder performs encoding processing on the image based on the target image texture data after the image quality enhancement processing.
12. An image processing device implemented by interacting with a cloud server, comprising: A collection component injection module is used to obtain a picture collection component provided by a cloud server and inject the picture collection component into an image rendering engine of an image loading program; A texture data providing module, used for capturing the image rendered by the image rendering engine through the hook function of the picture acquisition component, and providing the initial image texture data corresponding to the image to the cloud server, so that the cloud server can perform image encoding; The cloud server converts the initial image texture data into target image texture data suitable for the processor computing architecture; the target image texture data is subjected to image quality enhancement processing using the image quality processing algorithm of the processor computing architecture; The target image texture data after the image quality enhancement processing is sent to the encoder, so that the encoder performs encoding processing on the image based on the target image texture data after the image quality enhancement processing.
13. An electronic device comprising a memory, a processor and a computer program stored in the memory, wherein the processor implements the method according to any one of claims 1 to 10 when executing the computer program.
14. A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.
Citation Information
Patent Citations
Video encoding method and device and storage medium
CN111093096A
Cloud game implementation method and device, medium and electronic equipment
CN113015021A
Coded image quality adjusting method and device
CN114554204A
Image quality processing method and device, electronic equipment and storage medium
CN116883564A
Image transmission method for movable platform, movable platform, and readable storage medium
WO2022217557A1
