Cross-device adaptive graphical display control method and device, equipment and medium

Abstract

The present application relates to the technical field of computer graphics processing, and provides a cross-device adaptive graphics display control method and device, equipment and medium, which realizes device capability perception through the combination of active detection and passive reporting, and supplements incomplete capability information by using a device fingerprint library; a deep learning model is used to extract features from input original graphics data, a multi-dimensional adaptive decision model is constructed based on obtained device capability information and a standardized feature vector, a multi-objective optimization algorithm is used to output a rendering parameter adjustment scheme, original graphics data is rendered in real time according to the rendering parameter adjustment scheme, the display output of processed graphics is collected and the display effect is verified, and readjustment is returned when the verification fails, so that the dynamic balance between graphics display effect and resource occupation in a cross-device scene is realized, the cross-device color display consistency is ensured, and the rendering strategy is dynamically adjusted in real time in response to device state changes.

Description

Technical Field

[0001] This invention belongs to the field of computer graphics processing technology, and relates to a cross-device adaptive graphics display control method, device, equipment and medium. Background Technology

[0002] With the development of IoT technology, display devices are becoming increasingly diversified, including smartphones, tablets, smart cockpit displays, industrial control panels, AR devices, and VR devices. These devices differ significantly in hardware parameters and software environments. Hardware parameters include display resolution, screen size, pixel density, memory capacity, GPU model, and rendering capabilities; software environments include operating system type and version, supported graphics API versions, and display driver models.

[0003] Currently, the basic process of graphics display control involves acquiring graphics data, calling graphics rendering interfaces, and outputting the display. Common graphics rendering interfaces include OpenGL ES, Vulkan, and Direct3D, with different interfaces suitable for different operating systems and hardware platforms. Regarding color spaces, mainstream standards include sRGB, DCI-P3, and Rec.2020, with different devices supporting different standards. As for texture compression formats, mainstream formats include ETC2, ASTC, and PVRTC. ETC2 is suitable for OpenGL ES devices, ASTC for ARM architecture devices, and PVRTC for PowerVR GPU devices.

[0004] Existing graphics display control technologies typically employ fixed rendering parameters. This means a set of rendering parameters is pre-configured for a specific device, and these fixed parameters are used directly for graphics rendering at runtime. Specifically, during application development, rendering parameters, including resolution, texture compression format, and color space, are manually configured for the target device. At runtime, the pre-configured fixed parameters are used to call the graphics rendering interface for rendering. If adaptation to other devices is required, a separate adaptation version must be developed for each device, and the rendering parameters must be reconfigured. Some existing solutions use a simple resolution scaling strategy. When the target device resolution differs from the original graphics resolution, linear scaling is used to adjust the graphics resolution, but this does not consider edge features, easily resulting in jagged edges and blurring. Therefore, these traditional technologies still suffer from poor adaptability to cross-device scenarios, unreasonable resource consumption, poor color consistency, and lag in response, resulting in a comprehensive performance problem. Summary of the Invention

[0005] To address the problems existing in the aforementioned traditional technologies, this invention proposes a cross-device adaptive graphics display control method, a cross-device adaptive graphics display control device, a computer device, and a computer-readable storage medium, which can solve the comprehensive performance problems of poor graphics display adaptability, unreasonable resource consumption, poor color consistency, and lag in cross-device scenarios.

[0006] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions: On the one hand, a cross-device adaptive graphics display control method is provided, including the following steps: By actively probing and passively reporting, the hardware parameters and software environment information of the target display device are obtained, and the device capability information of the target display device is determined. The incomplete device capability information is supplemented by a pre-built device fingerprint database. The device fingerprint database contains parameter templates of mainstream display devices. A deep learning model is used to extract features from the raw graphic data of the input target display device to obtain a standardized feature vector. The standardized feature vector includes graphic type, texture complexity, color space type, and dynamic feature information. Based on device capability information and standardized feature vectors, a multi-dimensional adaptation decision model is constructed and a multi-objective optimization algorithm is used to output a rendering parameter adjustment scheme. The rendering parameter adjustment scheme includes resolution scaling ratio, texture compression level, color space conversion rules, and rendering level simplification strategy. Based on the rendering parameter adjustment scheme, the graphics processing interface of the target display device is called to perform real-time rendering processing on the original graphics data. Collect the display output data of the rendered graphics and verify the display effect; if the verification fails, return to the process of building a multi-dimensional adaptation decision model and readjust the rendering parameter adjustment scheme.

[0007] In one embodiment, the process of obtaining the hardware parameters and software environment information of the target display device through active detection and passive reporting includes: Send a standardized query command to the target display device to obtain the hardware parameters of the target display device; the hardware parameters include display resolution, screen size, pixel density, memory capacity, GPU model and rendering capabilities; Obtain software environment information through the system interface of the target display device; software environment information includes operating system type and version, graphics API supported version, and display driver model; The process of supplementing incomplete equipment capability information using a pre-built equipment fingerprint database includes: When the device capability information obtained through active detection or passive reporting is incomplete, the missing information is supplemented by parameter templates in the device fingerprint database; the device fingerprint database matches basic parameters with the device model, and the parameter templates in the device fingerprint database are updated through the cloud server.

[0008] In one embodiment, the step of using a deep learning model to extract features from the raw graphic data of the input target display device to obtain a standardized feature vector includes: The graphic type is identified based on the original graphic data; graphic types include 2D vector graphics, 2D bitmap graphics, 3D model graphics, and dynamic video frames; Texture complexity is quantified into multiple levels based on the quantity and density of detailed textures in the graphics; each level includes 1 to 10 levels indicating from simple to complex. Color space parameters are extracted from the original graphic data; the color space parameters include the color space type and RGB channel variance of the original graphic data. Dynamic features are extracted from dynamic video frames in the original graphic data; dynamic features include frame rate requirements and the amplitude of screen motion.

[0009] In one embodiment, the resolution scaling ratio is determined using a non-linear scaling algorithm based on the display resolution of the target display device and the original resolution of the original graphics data. The texture compression level is determined based on the compression format supported by the target display device's GPU and the texture complexity; the texture compression level includes 1 to 5 levels from no compression to the highest compression ratio, and matches the optimal compression format; Color space conversion rules include determining the conversion matrix between the color space of the original graphic data and the color space supported by the target display device; The rendering layer simplification strategy includes reducing the number of polygons in the model and reducing the complexity of lighting effects for 3D graphics based on the GPU rendering capabilities of the target display device.

[0010] In one embodiment, the step of calling the graphics processing interface of the target display device to perform real-time rendering processing of the original graphics data according to the rendering parameter adjustment scheme includes: Based on the resolution scaling ratio, the graphics resolution is adjusted using a non-linear scaling algorithm through the scaling function of the graphics processing interface, and the resolution scaling ratio is adaptively adjusted based on the graphics edge features. Based on the texture compression level and compression format, the GPU hardware compression interface is invoked to compress the texture data; Color space conversion is performed through color matrix multiplication operations; the color space conversion includes dynamic switching between sRGB, DCI-P3 and Rec.2020; Rendering layers are simplified using a 3D model simplification algorithm; The device operation status of the target display device is monitored in real time at preset time intervals; when the load exceeds the preset upper limit threshold, the rendering complexity is dynamically reduced, and when the load is below the preset lower limit threshold, the display effect is improved. The load includes CPU load and GPU load.

[0011] In one embodiment, the objective function of the multi-objective optimization algorithm is: Min F = α × (resource utilization rate) - β × (display performance score) Wherein, resource utilization rate = w1 × CPU utilization rate + w2 × GPU utilization rate + w3 × memory utilization rate, where w1 is the weighting coefficient of CPU utilization rate, w2 is the weighting coefficient of GPU utilization rate, and w3 is the weighting coefficient of memory utilization rate. Display performance score = w4 × Clarity score + w5 × Color reproduction score + w6 × Frame rate stability score, where w4 is the weighting coefficient for clarity, w5 is the weighting coefficient for color reproduction, w6 is the weighting coefficient for frame rate stability, α is the target weighting coefficient for resource utilization, and β is the target weighting coefficient for display performance score.

[0012] In one embodiment, the step of acquiring the display output data of the rendered graphics and verifying the display effect includes: The rendered graphics are captured using a screen capture device, and machine vision inspection metrics are calculated. These metrics include sharpness, color fidelity, and frame rate stability. Sharpness is calculated using peak signal-to-noise ratio (PSNR), with a preset threshold of at least 35 dB. Color fidelity is calculated by comparing the color deviation between the output image and the original image, with a preset threshold of at least 2. Frame rate stability is calculated by analyzing the actual displayed frame rate and determining the frame rate fluctuation, which is the difference between the maximum and minimum values ​​of the actual displayed frame rate. The preset frame rate stability threshold is at least 5%. The system obtains user ratings for the display effect through the configured user rating interface and calculates the average user rating. If the machine vision detection indicators do not meet the preset thresholds or the average user score is lower than the preset score threshold, the verification is deemed to have failed.

[0013] On the other hand, a cross-device adaptive graphics display control device is also provided, comprising: The device capability perception module is used to obtain the hardware parameters and software environment information of the target display device through active detection and passive reporting, and then determine the device capability information of the target display device. It uses a pre-built device fingerprint database to supplement incomplete device capability information; the device fingerprint database contains parameter templates of mainstream display devices. The graphics feature extraction module is used to extract features from the raw graphics data of the input target display device using a deep learning model to obtain standardized feature vectors. The standardized feature vectors include graphics type, texture complexity, color space type, and dynamic feature information. The adaptation decision module is used to construct a multi-dimensional adaptation decision model based on device capability information and standardized feature vectors, and to output a rendering parameter adjustment scheme using a multi-objective optimization algorithm. The rendering parameter adjustment scheme includes resolution scaling ratio, texture compression level, color space conversion rules, and rendering level simplification strategy. The real-time rendering module is used to adjust the scheme according to the rendering parameters and call the graphics processing interface of the target display device to perform real-time rendering processing of the original graphic data. The display verification module is used to collect the display output data of the rendered graphics and verify the display effect. If the verification fails, it returns to the process of building a multi-dimensional adaptation decision model and readjusts the rendering parameter adjustment scheme.

[0014] In another aspect, a computer device is also provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of the above-described cross-device adaptive graphics display control method.

[0015] Furthermore, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the aforementioned cross-device adaptive graphics display control method.

[0016] One of the above technical solutions has the following advantages and beneficial effects: The aforementioned cross-device adaptive graphics display control method, device, equipment, and medium achieve comprehensive acquisition of cross-device hardware parameters and software environment information through a device capability perception mechanism combining active detection and passive reporting, and a device fingerprint database. It achieves automatic parsing of graphics features through a deep learning model, achieves dynamic balance between display effect and resource consumption through a multi-objective optimization algorithm, and achieves real-time adjustment of rendering strategy through real-time status monitoring and dynamic fine-tuning mechanism. This solves the problems of poor graphics display adaptability, unreasonable resource consumption, poor color consistency, and response lag in cross-device scenarios. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a cross-device adaptive graphics display control method in one embodiment; Figure 2 This is a schematic diagram of the data flow for device capability awareness in one embodiment; Figure 3 This is a schematic diagram of the feature extraction process of a deep learning model in one embodiment. Figure 4 This is a schematic diagram illustrating the input-output relationship of a multi-objective optimization algorithm in one embodiment; Figure 5 This is a flowchart illustrating the dynamically fine-tuned judgment logic in one embodiment; Figure 6 This is a flowchart illustrating the verification and feedback logic in one embodiment; Figure 7 This is a schematic diagram of the overall application flow of a cross-device adaptive graphics display control method in one embodiment; Figure 8 This is a schematic diagram of the module architecture of a cross-device adaptive graphics display control device in one embodiment. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention.

[0020] It should be noted that, in this document, the reference to "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The presentation of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will understand that the embodiments described herein can be combined with other embodiments. The term "and / or" as used herein refers to any combination of one or more of the associated listed items, and all possible combinations, including such combinations.

[0021] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0022] Traditional methods are often developed for specific devices, lacking a universal adaptation mechanism, resulting in significant differences in the display of the same graphics across different devices. Low-resolution graphics appear blurry on high-resolution devices, while high-resolution graphics experience frame rate stuttering on low-performance devices. Each new device type requires a separate adapted version, leading to high development costs. Fixed rendering parameters cannot be dynamically adjusted based on the actual hardware capabilities of the device. High-performance devices do not fully utilize their hardware capabilities, resulting in suboptimal display effects; low-performance devices experience stuttering and frame rate drops due to excessive rendering load. Different devices use different color space standards (e.g., mobile phones support sRGB, while smart cockpits support DCI-P3), lacking a dynamic conversion mechanism, resulting in low color fidelity for the same graphics on different devices, with color deviation values ​​often exceeding 5. Furthermore, it cannot detect changes in device operating status in real time. When device load fluctuates, such as a sudden increase in GPU load due to a new application launching in the background, the rendering strategy cannot be adjusted in time, leading to a sharp drop in frame rate or screen stuttering.

[0023] Therefore, how to achieve adaptive adaptation of graphics display across different devices, so that display devices with different hardware parameters and software environments can achieve a balance between graphics display effect and resource consumption, while ensuring consistent color display across devices and dynamically adjusting rendering strategies in real time in response to changes in device status, has become a technical problem that urgently needs to be solved in this field.

[0024] In one embodiment, such as Figure 1 As shown, a cross-device adaptive graphics display control method is provided, which may include the following processing steps S10 to S18: S10: After obtaining the hardware parameters and software environment information of the target display device through active detection and passive reporting, the device capability information of the target display device is determined, and the incomplete device capability information is supplemented by a pre-built device fingerprint database; wherein, the device fingerprint database contains parameter templates of mainstream display devices.

[0025] Understandably, in existing technologies, developers need to configure rendering parameters such as resolution, texture compression format, and color space for each target device during the application development phase. When facing new device models or devices without pre-configuration, a new adapted version needs to be developed, resulting in high adaptation costs and narrow coverage. This step combines active detection and passive reporting mechanisms to simultaneously obtain device information from both the hardware and software levels. Active detection directly obtains hardware parameters by sending standardized query commands to the target device, while passive reporting reads software environment information by calling the target device's system interface. The two mechanisms complement and verify each other, ensuring the completeness and accuracy of the information. Target display devices include, but are not limited to, smartphones, tablets, smart cockpit displays, industrial control panels, AR devices, and VR devices.

[0026] Furthermore, by introducing a pre-built device fingerprint database, when the information obtained through active detection and passive reporting is incomplete, the missing information can be quickly supplemented using the parameter templates pre-stored in the fingerprint database. This allows this embodiment to maintain normal operation even when device information is incomplete. Active detection directly acquires dynamic hardware parameters, passive reporting supplements static software environment information, and the fingerprint database provides a fallback for the former two. This enables the method to support multiple architectures such as ARM and RISC-V / x86, as well as multiple operating systems such as Android, iOS, and Windows / embedded Linux, resulting in high compatibility coverage. The device capability information obtained in step S10 provides an accurate and comprehensive input basis for subsequent multi-dimensional adaptation decisions and is a prerequisite for all subsequent adaptive processing.

[0027] S12 uses a deep learning model to extract features from the raw graphic data of the input target display device to obtain a standardized feature vector; the standardized feature vector includes graphic type, texture complexity, color space type and dynamic feature information.

[0028] It is understandable that in existing technologies, developers need to manually judge the complexity of graphics and configure corresponding rendering strategies based on experience for different types of graphics data, lacking automated and standardized feature quantization methods. This embodiment uses a deep learning model built on a CNN convolutional neural network. This model has been trained and optimized with a large number of graphics samples and can automatically identify the graphics type of the input raw graphics data, including 2D vector graphics, 2D bitmaps, 3D model graphics, and dynamic video frames.

[0029] Meanwhile, the model quantifies texture complexity into standardized levels (e.g., 1 to 10) based on the quantity and density of detailed textures in the image, transforming texture complexity from a subjective judgment into a quantifiable objective indicator. It extracts color space parameters to obtain the color space type and RGB channel variance of the original image, reflecting color richness. For dynamic video frames, it also extracts dynamic features such as frame rate requirements and motion amplitude. These four types of feature information—image type recognition, texture complexity quantification, color space parameter extraction, and dynamic feature extraction—characterize the content characteristics of the original image from different dimensions. Image type determines which features are meaningful, texture complexity influences the selection of subsequent compression strategies, color space parameters guide the direction of color conversion, and dynamic features provide temporal constraints for video frame processing. Together, these four elements constitute a standardized feature vector, providing complete and structured input data for the subsequent multi-dimensional adaptation decision model, enabling fine-tuning of adaptation decisions based on the specific content characteristics of the image.

[0030] S14, based on device capability information and standardized feature vectors, constructs a multi-dimensional adaptation decision model and uses a multi-objective optimization algorithm to output a rendering parameter adjustment scheme; the rendering parameter adjustment scheme includes resolution scaling ratio, texture compression level, color space conversion rules, and rendering layer simplification strategy.

[0031] It is understandable that in existing technologies, rendering parameters are pre-configured fixed values, which cannot be dynamically adjusted according to the actual capabilities of the device and the characteristics of the graphics content. This leads to resource waste on high-performance devices and rendering overload on low-performance devices. Here, we use the acquired device capability information and standardized feature vectors as input to construct a multi-dimensional adaptation decision model. This model considers both the constraints on the device side and the content characteristics on the graphics side, and makes joint decisions from four dimensions: resolution, texture, color, and rendering complexity.

[0032] The resolution scaling factor comprehensively considers both the device's display resolution and the original graphics resolution, ensuring that the scaled graphics match the device's physical display capabilities while preserving as much detail as possible. The texture compression level is determined based on the compression formats supported by the GPU and the complexity of the graphics textures, striking a balance between compression efficiency and image quality loss. Color space conversion rules determine the conversion matrix from the original color space to the device's color space, ensuring consistent color display across devices. The rendering level simplification strategy targets 3D graphics, determining the degree of model simplification based on the GPU's rendering capabilities. These four dimensions are not independent decisions but are coupled and synergistically optimized: for example, when the resolution scaling factor is large, the texture compression level can be appropriately increased without significantly affecting visual effects; when the device's GPU capabilities are weak, the aggressiveness of texture compression can be reduced while simplifying the rendering levels to retain more visual detail. By employing a multi-objective optimization algorithm, using a weighted balance between resource utilization and display effect score as the optimization objective, a coordinated and unified rendering parameter adjustment scheme is output.

[0033] S16, based on the rendering parameter adjustment scheme, calls the graphics processing interface of the target display device to perform real-time rendering processing of the original graphics data.

[0034] It is understandable that in existing technologies, rendering processing uses fixed parameters configured once, making it impossible to adjust in a timely manner when the device's operating state changes. Here, the rendering parameter adjustment scheme output in the previous steps is transformed into actual graphics rendering operations, and the specific rendering processing is executed by calling the graphics processing interface.

[0035] Specifically, in terms of resolution adjustment, the graphics resolution is adjusted by using a non-linear scaling algorithm through the scaling function of the graphics processing interface. This non-linear scaling algorithm adaptively adjusts the scaling ratio based on the edge features of the graphics, applying a smaller scaling deformation in the edge area to maintain sharpness, and applying a larger scaling deformation in the flat area to improve efficiency, thereby effectively avoiding the jagged distortion caused by traditional linear scaling.

[0036] In terms of texture compression, the GPU hardware compression interface is called to compress texture data according to compression level and format. It supports adaptive selection from multiple compression formats such as ETC2, ASTC and PVRTC, and automatically matches the optimal compression format based on the device GPU's supported type.

[0037] In terms of color space conversion, dynamic switching between sRGB, DCI-P3, and Rec.2020 is achieved through color matrix multiplication, with a conversion matrix (e.g., 3×3) precisely mapping each color channel. Regarding rendering layer simplification, redundant polygons are removed using a 3D model simplification algorithm, and lighting rendering parameters are adjusted. These four rendering processes are closely coordinated: resolution adjustment changes the pixel sampling range, directly affecting the visual effect of subsequent texture compression; color space conversion changes the representation of pixel color values, affecting the color accuracy of the final rendered output; and model simplification changes the number of geometry objects to be rendered, affecting GPU load allocation. These four processes execute collaboratively, jointly determining the final rendering quality and resource consumption level.

[0038] In addition, this step may optionally include a dynamic fine-tuning mechanism: for example, monitoring the device's operating status in real time at preset time intervals; when the CPU or GPU load exceeds a preset upper limit threshold, dynamically reducing rendering complexity, such as increasing texture compression levels to further simplify 3D model rendering layers; when the load is below a preset lower limit threshold, improving display effects, such as increasing rendering resolution and adding texture details. This dynamic fine-tuning mechanism can sense changes in the device's operating status in real time and adjust rendering strategies promptly to adapt to load fluctuations, resulting in low response latency and fundamentally solving the problem of sudden frame rate drops caused by device load fluctuations in existing technologies.

[0039] S18: Collect the display output data of the rendered graphics and verify the display effect; if the verification fails, return to the process of building a multi-dimensional adaptation decision model and readjust the rendering parameter adjustment scheme.

[0040] It is understandable that in existing technologies, rendering parameters are fixed after configuration, lacking verification and feedback correction of actual display effects. Here, verification is performed by collecting and processing the display output data of the graphics, constructing a complete closed-loop optimization circuit. In terms of machine vision inspection, the display output image of the processed graphics is acquired through a screen capture device, and machine vision inspection indicators are calculated. For example, peak signal-to-noise ratio (PSNR) is used to calculate the sharpness indicator; PSNR is an objective indicator of image quality, with higher values ​​indicating better image quality. The color deviation value ΔE between the output image and the original image is calculated as a color reproduction indicator; ΔE measures the difference between two colors. The actual display frame rate is statistically analyzed, and the frame rate fluctuation value is calculated as a frame rate stability indicator.

[0041] Optionally, a user rating interface can be provided for user experience feedback, allowing users to rate the display effect. The system collects user ratings and calculates the average user rating. When the machine vision inspection indicators fail to meet the preset threshold, or the average user rating is lower than the preset threshold, the verification is deemed unsuccessful, and the system automatically returns to step S14 to build a multi-dimensional adaptation decision model, readjusting the rendering parameter adjustment scheme. Machine vision inspection provides an objective and quantifiable evaluation of the display effect, while user experience feedback provides a subjective evaluation dimension. The two complement each other and together constitute a complete verification system. This step forms a feedback loop with step S14, thus possessing self-optimization capabilities and gradually approaching the optimal rendering parameter configuration through iterative adjustments.

[0042] The aforementioned cross-device adaptive graphics display control method achieves comprehensive acquisition of cross-device hardware parameters and software environment information through a device capability perception mechanism that combines active detection and passive reporting, and a device fingerprint database. It also achieves automatic parsing of graphics features through a deep learning model, dynamic balance between display effect and resource consumption through a multi-objective optimization algorithm, and real-time adjustment of rendering strategy through real-time status monitoring and dynamic fine-tuning mechanism. This solves the problems of poor graphics display adaptability, unreasonable resource consumption, poor color consistency, and response lag in cross-device scenarios.

[0043] In one embodiment, the process of obtaining the hardware parameters and software environment information of the target display device through active detection and passive reporting in step S10 may specifically include the following processing: Send a standardized query command to the target display device to obtain the hardware parameters of the target display device; the hardware parameters include display resolution, screen size, pixel density, memory capacity, GPU model and rendering capabilities; Obtain software environment information through the system interface of the target display device; the software environment information includes operating system type and version, graphics API supported version, and display driver model.

[0044] Step S10 above involves supplementing incomplete device capability information using a pre-built device fingerprint database. Specifically, this process may include the following steps: When the device capability information obtained through active detection or passive reporting is incomplete, the missing information is supplemented by parameter templates in the device fingerprint database; the device fingerprint database matches basic parameters with the device model, and the parameter templates in the device fingerprint database are updated through the cloud server.

[0045] Understandable, such as Figure 2 As shown, in terms of active detection, standardized query commands are sent to the target display device. These commands are sent via HDMI-CEC, USB HID, and network protocol commands to obtain hardware parameters. Specific hardware parameters include: display resolution (1920×1080, 2560×1440), screen size (e.g., 5.5 inches, 15.6 inches), pixel density (PPI, pixels per inch), GPU model and rendering capability (measured in GFLOPS), and memory capacity. The advantage of active detection is that it can obtain the device's current hardware status in real time, providing timely information.

[0046] In passive reporting, software environment information is obtained through the system interfaces of the target display device (such as Android's DisplayManager API, Windows' Direct3D API, and embedded Linux's DRM API). The system interfaces used differ depending on the operating system: Android devices use the DisplayManager interface; Windows devices use the Direct3D interface; and embedded Linux devices use the DRM (Direct Rendering Manager) interface. The software environment information specifically includes the operating system type and version, the supported graphics API version (such as OpenGL ES 3.2 or Vulkan 1.1), and the display driver model. The advantage of passive reporting is that it can obtain system-level static configuration information, which is authoritative.

[0047] The device fingerprint database pre-stores a large number of parameter templates (e.g., 1000+) of mainstream display devices, supporting quick matching of basic parameters by device model. It also allows users to add custom parameters for specific devices. The parameter templates in the device fingerprint database are updated regularly via a cloud server. When information obtained through active detection or passive reporting is incomplete—for example, if some devices do not support standardized query commands, or if system interface permissions are limited and complete software environment information cannot be obtained—the parameter templates in the device fingerprint database can supplement the missing information.

[0048] The aforementioned active detection acquires the device's current real-time hardware status, passive reporting acquires the system-level static software configuration, and the device fingerprint database provides historically accumulated prior knowledge. While these three mechanisms differ in their information sources, coverage dimensions, and timeliness, they share the same goal: to comprehensively and accurately acquire the target display device's capability information. When information acquisition through any one mechanism is limited, other mechanisms can provide supplementary or alternative mechanisms, ensuring the robustness and completeness of device capability perception. The device capability information acquired based on this mechanism provides complete device-side constraints for subsequent multi-dimensional adaptation decision models, forming the foundation for adaptive graphics display control.

[0049] In one embodiment, step S12 above may further include the following processing: The graphic type is identified based on the original graphic data; graphic types include 2D vector graphics, 2D bitmap graphics, 3D model graphics, and dynamic video frames; Texture complexity is quantified into multiple levels based on the quantity and density of detailed textures in the graphics; each level includes 1 to 10 levels indicating from simple to complex. Color space parameters are extracted from the original graphic data; the color space parameters include the color space type and RGB channel variance of the original graphic data. Dynamic features are extracted from dynamic video frames in the original graphic data; dynamic features include frame rate requirements and the amplitude of screen motion.

[0050] Understandable, such as Figure 3 As shown, the deep learning model is built on a CNN convolutional neural network. This model is trained and optimized through a large number of image samples, such as more than 100,000 image samples, enabling multi-dimensional feature extraction from the input raw image data. The input format of the raw image data supports various types, including image formats such as PNG and JPG, SVG vector graphics format, 3D model formats (such as OBJ and FBX), and dynamic video frames (such as YUV and RGB formats).

[0051] Feature extraction begins with image type identification, distinguishing which of the following four types the input raw image data belongs to: 2D vector graphics, 2D bitmap graphics, 3D model images, or dynamic video frames. Correct image type identification is a prerequisite for subsequent feature extraction strategy selection, as different types of images differ in their focus and methods for feature extraction.

[0052] Building upon this foundation, texture complexity is quantified: based on the quantity and density of detailed textures in the image, texture complexity is quantified into multiple levels, such as 1 to 10, where level 1 represents the simplest texture and level 10 represents the most complex texture. Quantifying texture complexity transforms the originally subjective judgment of complexity into a comparable and calculable objective indicator, providing a quantitative basis for the subsequent selection of texture compression levels.

[0053] Simultaneously, color space parameters are extracted, including the color space type and RGB channel variance of the original graphic data. Color space types can include sRGB, DCI-P3, and Rec.2020, etc. The RGB channel variance reflects the dispersion of color distribution in the graphic; the larger the variance, the richer the colors.

[0054] For dynamic video frames, additional dynamic features can be extracted, including frame rate requirements and motion amplitude. Frame rate requirements specify the expected playback frame rate for the video frame, such as 30 FPS or 60 FPS; motion amplitude reflects the drastic changes in content between consecutive frames, with a larger motion amplitude indicating higher dynamic complexity.

[0055] The feature extraction of the above four dimensions—graphic type recognition, texture complexity quantification, color space parameter extraction, and dynamic feature extraction—together constitutes a complete feature description system. Graphic type determines which feature dimensions are meaningful; for example, dynamic features are only effective for dynamic video frames. Texture complexity quantification and color space parameter extraction characterize graphic content from structural and color dimensions, respectively. Graphic type, in turn, provides context for interpreting texture complexity and color parameters. This feature description system unifies the scattered and heterogeneous information in the original graphic data into standardized feature vectors, enabling downstream adaptation decision models to perform reasoning and decision-making based on a unified data structure.

[0056] In one embodiment, the resolution scaling ratio is determined using a non-linear scaling algorithm based on the display resolution of the target display device and the original resolution of the original graphics data. The texture compression level is determined based on the compression format supported by the target display device's GPU and the texture complexity; the texture compression level includes levels 1 to 5, ranging from no compression to the highest compression ratio, and matches the optimal compression format. The color space conversion rules include determining the conversion matrix between the color space of the original graphics data and the color space supported by the target display device. The rendering level simplification strategy includes reducing the number of polygons in the model and lowering the complexity of lighting effects based on the GPU rendering capabilities of the target display device for 3D graphics.

[0057] Understandable, such as Figure 4 As shown, the rendering parameter adjustment scheme output by the multi-dimensional adaptation decision model includes adjustment decisions in four dimensions: The first dimension is the resolution scaling ratio. This scaling ratio is determined using a non-linear scaling algorithm based on the target display device's display resolution and the original resolution of the original graphics data. Non-linear scaling algorithms, such as Lanczos interpolation, are implemented using the Lanczos kernel function and can better preserve image edge details compared to traditional bilinear interpolation. Specifically, when the device's display resolution is higher than the original graphics resolution, the scaling ratio is greater than 1, requiring upsampling and magnification. In this case, the non-linear scaling algorithm applies a smaller scaling distortion to edge areas to maintain sharpness. When the device's display resolution is lower than the original graphics resolution, the scaling ratio is less than 1, requiring downsampling and shrinking. In this case, the non-linear scaling algorithm applies a larger scaling distortion to flat areas to improve efficiency.

[0058] The second dimension is the texture compression level. The texture compression level is determined by the compression format supported by the target display device's GPU and the complexity of the graphics texture. Texture compression levels are divided into 1 to 5, where level 1 represents no compression, suitable for scenarios with extremely low texture complexity or sufficient device GPU performance; level 5 represents the highest compression ratio, suitable for scenarios with high texture complexity and limited device GPU performance. In addition to determining the compression level, the optimal compression format must also be matched. The optimal compression format is selected from ETC2, ASTC, and PVRTC. ETC2 is the OpenGL ES standard format, ASTC is a high-quality compression format primarily used by ARM, and PVRTC is a PowerVR GPU-specific format. The choice of compression format is entirely based on the hardware support type of the device's GPU, prioritizing compression formats natively supported by the GPU to leverage hardware acceleration advantages.

[0059] The third dimension is the color space conversion rule. This rule specifically determines the conversion matrix between the color space of the original graphic data and the color space supported by the target display device. For example, when the original graphic uses the sRGB color space and the target device supports the DCI-P3 color space, a 3×3 conversion matrix is ​​constructed. Through color matrix multiplication, the values ​​of each sRGB channel are precisely mapped to the values ​​of each DCI-P3 channel, achieving accurate color conversion. sRGB is a universal display standard, DCI-P3 is a digital cinema standard, and Rec.2020 is an ultra-high-definition television standard; conversions between these three are all accomplished through corresponding 3×3 conversion matrices.

[0060] The fourth dimension is the rendering layer simplification strategy. This strategy targets 3D graphics and determines the degree of model simplification based on the GPU rendering capabilities of the target display device. Specifically, it includes reducing the number of polygons in the model and decreasing the complexity of lighting effects. When the GPU rendering capability is strong, a larger number of polygons and higher lighting effect complexity are retained to ensure image quality; when the GPU rendering capability is weak, the number of polygons is reduced and the complexity of lighting effects is decreased to alleviate the rendering burden.

[0061] The decisions made in these four dimensions are not independent but rather closely coupled and synergistic. When the resolution scaling is large, the texture compression level can be appropriately increased without significantly impacting the visual experience due to the loss of image detail. When color space conversion introduces a significant loss in color gamut mapping, the texture compression level can be appropriately reduced to retain more color accuracy information. When the model rendering layers are significantly simplified, the resolution scaling can be appropriately increased to maintain the overall image quality. The multi-dimensional adaptation decision model considers these four dimensions simultaneously through a multi-objective optimization algorithm, ensuring the coordination and optimality of the overall rendering parameter adjustment scheme.

[0062] In one embodiment, the objective function of the multi-objective optimization algorithm is: Min F = α × (resource utilization rate) - β × (display performance score) Wherein, resource utilization rate = w1 × CPU utilization rate + w2 × GPU utilization rate + w3 × memory utilization rate, where w1 is the weighting coefficient of CPU utilization rate, w2 is the weighting coefficient of GPU utilization rate, and w3 is the weighting coefficient of memory utilization rate. Different application scenarios can adjust the values ​​of w1, w2, and w3 according to the different sensitivities to various resources, and the sum of w1, w2, and w3 is 1.

[0063] Display performance score = w4 × Clarity score + w5 × Color reproduction score + w6 × Frame rate stability score, where w4 is the weighting coefficient for clarity, w5 is the weighting coefficient for color reproduction, and w6 is the weighting coefficient for frame rate stability. The sum of w4, w5, and w6 is 1. α is the target weighting coefficient for resource utilization, and β is the target weighting coefficient for the display performance score. α ∈ [0.3, 0.7], β ∈ [0.3, 0.7], and can be dynamically adjusted according to the application scenario. In gaming scenarios, where display performance is critical, a larger value for β, such as 0.7, is set to prioritize ensuring good display quality. In industrial control scenarios, where resource utilization is critical, a larger value for α, such as 0.7, is set to prioritize reducing resource consumption.

[0064] It is understandable that this objective function adopts a minimization optimization paradigm, where a smaller value of function F indicates better overall performance. The product of α and resource utilization reflects the penalty for resource consumption, while the product of β and display effect score reflects the reward for display quality. Together, they enable the optimization algorithm to seek the optimal balance between resource consumption and display effect. The dynamic adjustability of the objective weight coefficients α and β allows the above method to flexibly adapt to the differentiated needs of different application scenarios, rather than being limited to a single fixed strategy.

[0065] The sub-weight coefficients w1, w2, and w3 for resource utilization, and the sub-weight coefficients w4, w5, and w6 for display effect scoring, also exhibit a synergistic relationship: for example, when the CPU is found to be the bottleneck resource during actual operation, the weight of w1 can be appropriately increased to enhance the optimization algorithm's sensitivity to CPU utilization; when the application scenario has strict requirements for color accuracy, the weight of w5 can be appropriately increased to enhance the emphasis on color reproduction. This multi-level weight system design allows the optimization algorithm to make fine-grained trade-offs between different resources and different quality dimensions.

[0066] Furthermore, step S16 above may further include the following processing: Based on the resolution scaling ratio, the graphics resolution is adjusted using a non-linear scaling algorithm through the scaling function of the graphics processing interface, and the resolution scaling ratio is adaptively adjusted based on the graphics edge features. Based on the texture compression level and compression format, the GPU hardware compression interface is invoked to compress the texture data; Color space conversion is performed through color matrix multiplication operations; the color space conversion includes dynamic switching between sRGB, DCI-P3 and Rec.2020; Rendering layers are simplified using a 3D model simplification algorithm; The device operation status of the target display device is monitored in real time at preset time intervals; when the load exceeds the preset upper limit threshold, the rendering complexity is dynamically reduced, and when the load is below the preset lower limit threshold, the display effect is improved. The load includes CPU load and GPU load.

[0067] It is understandable that, regarding resolution adjustment, the graphics resolution is adjusted through a scaling function in the graphics processing interface based on the resolution scaling ratio output in step S14. This scaling function employs a non-linear scaling algorithm, adaptively adjusting the scaling ratio according to the edge features of the graphics. Specifically, an edge detection algorithm first identifies edge regions and flat regions in the graphics. For edge regions, a smaller scaling deformation is applied to maintain sharpness and avoid jagged edges; for flat regions, a larger scaling deformation can be applied to improve processing efficiency. This edge-adaptive non-linear scaling strategy makes the scaled graphics visually clearer and the edges sharper.

[0068] Regarding texture compression, based on the texture compression level and format output in step S14, the GPU hardware compression interface is invoked to compress the texture data. The compression process supports adaptive selection from multiple compression formats, including ETC2, ASTC, and PVRTC, based on the hardware support type of the device's GPU. If the device GPU is an ARM Mali series, the ASTC format is preferred; if the device GPU is a PowerVR series, the PVRTC format is preferred; and if the device GPU is a general-purpose OpenGL ES implementation, the ETC2 format is selected. This adaptive format selection mechanism based on GPU hardware capabilities fully leverages the hardware acceleration advantages of the GPU.

[0069] In terms of color space conversion, the conversion from the original color space to the device color space is achieved through color matrix multiplication. Specifically, the target conversion matrix, determined by the color space conversion rules, is multiplied with the pixel color values ​​of the original image, mapping the RGB three-channel values ​​in the original color space to the corresponding values ​​in the target color space. This method supports dynamic switching between sRGB, DCI-P3, and Rec.2020, ensuring that the same image displays consistent color effects on devices with different color space standards.

[0070] In terms of rendering level simplification, a 3D model simplification algorithm is used. This algorithm analyzes the mesh data of the 3D model, identifies and removes redundant polygons that contribute little to the visual effect, and adjusts the lighting rendering parameters to match the complexity of the simplified model.

[0071] like Figure 5 As shown, in terms of dynamic fine-tuning, the device's operating status is monitored in real time at preset time intervals (e.g., 100ms). The monitored objects include CPU load, GPU load, and memory usage. When the CPU load or GPU load exceeds a preset upper threshold (e.g., 80%), it indicates that the device is currently under high load. The rendering complexity is then dynamically reduced. This can specifically include increasing the texture compression level to reduce texture data transfer and processing overhead, and further simplifying the number of polygons in the 3D model to reduce vertex shading and rasterization load. When both CPU load and GPU load are below a preset lower threshold (e.g., 30%), it indicates that the device currently has sufficient rendering capacity. The display effect can then be improved. This can specifically include increasing the rendering resolution to increase image detail, and increasing texture detail to enrich the image content. This dynamic fine-tuning mechanism allows the rendering strategy to adapt to changes in the device's operating status in real time, automatically adjusting rendering parameters to maintain stable display effects when the device load fluctuates.

[0072] The five processes mentioned above—resolution adjustment, texture compression, color space conversion, rendering level simplification, and dynamic fine-tuning—constitute a complete system for real-time rendering control. The first four processes are positive rendering execution processes, while the fifth is a negative feedback adjustment process. Resolution adjustment determines the pixel sampling range, directly affecting the visual effect of texture compression and GPU load. Texture compression changes the transmission and processing overhead of texture data, affecting overall rendering efficiency. Color space conversion changes the final output value of pixels and is one of the final determining factors of display effect. Rendering level simplification directly changes the number of geometry objects the GPU needs to process and is the main means of adjusting GPU load. Dynamic fine-tuning continuously adjusts the first four processes based on the actual operating state to ensure that the rendering strategy is always in an optimal state.

[0073] In one embodiment, step S18 described above may further include the following processing: The rendered graphics are captured using a screen capture device, and machine vision inspection metrics are calculated. These metrics include sharpness, color fidelity, and frame rate stability. Sharpness is calculated using peak signal-to-noise ratio (PSNR), with a preset threshold of at least 35 dB. Color fidelity is calculated by comparing the output image with the original image, with a preset threshold of at least 2. Frame rate stability is calculated by determining the frame rate fluctuation after statistically analyzing the actual displayed frame rate; the frame rate fluctuation is the difference between the maximum and minimum actual displayed frame rate, with a preset threshold of at least 5%.

[0074] The system obtains user ratings for the display effect through the configured user rating interface and calculates the average user rating.

[0075] If the machine vision detection indicators do not meet the preset thresholds or the average user score is lower than the preset score threshold, the verification is deemed to have failed.

[0076] Understandable, such as Figure 6 As shown, in machine vision inspection, the processed graphic's display output image is acquired through a screen capture device, which includes a high-definition camera and a video capture card. Based on the acquired display output image, the following three machine vision inspection metrics are calculated: The first metric is the sharpness metric, calculated using PSNR (Peak Signal-to-Noise Ratio). PSNR is an objective indicator of image quality, measured in dB; a higher value indicates better image quality. The calculation compares the output image and the original image pixel-by-pixel, calculates the mean square error, and converts it to a PSNR value. The preset sharpness threshold is ≥35dB; a PSNR value below 35dB indicates insufficient image sharpness.

[0077] The second indicator is the color reproduction indicator, which calculates the color difference value ΔE between the output image and the original image. ΔE is an international standard indicator for measuring the difference between two colors, and is calculated using the CIEDE2000 color difference formula. ΔE less than or equal to 2 means that the human eye can hardly distinguish color differences. A preset threshold for the color reproduction indicator is ≤ 2, and when the ΔE value exceeds 2, it indicates that there is a perceptible color deviation.

[0078] The third indicator is the frame rate stability indicator, which counts the variation of the actual display frame rate and calculates the frame rate fluctuation value. The frame rate fluctuation value is the difference between the maximum value and the minimum value of the actual display frame rate, and the preset threshold for the frame rate stability indicator is ≤ 5%. When the frame rate fluctuation exceeds 5%, it indicates that there is obvious instability in the display process, and picture stuttering or frame dropping may occur.

[0079] In terms of user experience feedback, a user scoring interface can be provided for users to score the display effect, for example, a score from 1 to 5, where 5 is the best evaluation and 1 is the worst evaluation. The system collects scoring data from multiple users and calculates the average user score, which reflects the subjective satisfaction with the display effect.

[0080] When the machine vision detection indicator does not meet the preset threshold, or the average user score is lower than the preset score threshold (e.g., 3 points), it is determined that the verification fails, and a return mechanism is automatically triggered, returning to step S14 to rebuild the multi-dimensional adaptive decision model and adjust the rendering parameter adjustment scheme, and then re-execute the real-time rendering processing in step S16 and the verification in step S18 until the requirements are met, thereby forming a closed loop of iterative optimization.

[0081] The two verification methods of machine vision detection and user experience feedback form a complementary relationship between objective evaluation and subjective evaluation. Machine vision detection provides quantifiable and repeatable objective indicators, which is suitable for automated quality monitoring; user experience feedback provides subjective perception from end users, reflecting satisfaction in actual use scenarios. The two methods together form a complete display effect verification system, and failure of either party to meet the standard will trigger re-optimization, ensuring the rigor and comprehensiveness of the verification.

[0082] The three machine vision detection indicators, namely the sharpness indicator PSNR, the color reproduction indicator ΔE, and the frame rate stability indicator, comprehensively measure the display effect from three different dimensions: image quality, color accuracy, and timing stability. Among them, if PSNR meets the standard but ΔE exceeds the standard, it indicates that the picture is clear but the color is distorted; if ΔE meets the standard but the frame rate fluctuation exceeds the standard, it indicates that the color is accurate but the playback is not smooth. Only when all three indicators meet the standard at the same time can the comprehensive qualification of the display effect be ensured. The introduction of the average user score further compensates for the deviation between machine vision detection and user perception that may exist, making the verification result more practical.

[0083] The above-mentioned cross-device adaptive graphics display control method, such as Figure 7 As shown, through a device capability awareness mechanism that combines active detection and passive reporting, as well as a pre-built device fingerprint database, comprehensive acquisition of hardware parameters and software environment information of the target display device is achieved. It supports cross-device adaptation for multiple architecture devices such as ARM, RISC-V / x86, and multiple operating systems such as Android, iOS, Windows / embedded Linux, with a high adaptation coverage (≥95%). This replaces the method of manually configuring rendering parameters for each device in the existing technology, thereby achieving universal cross-device adaptation.

[0084] By employing a deep learning model built on a CNN convolutional neural network to extract features from raw graphic data, it automatically identifies 2D vector graphics, 2D bitmaps, 3D model images, and dynamic video frames. Texture complexity is quantified into 1 to 10 levels, outputting standardized feature vectors. This replaces the manual parameter configuration method used in existing technologies for different graphic types, achieving unified processing for various graphic types. A multi-dimensional adaptation decision model built through a multi-objective optimization algorithm aims at a weighted balance between resource utilization and display effect score. It supports dynamic adjustment of target weight coefficients based on application scenarios, replacing the fixed rendering parameter method of existing technologies. This reduces GPU utilization on low-performance devices by 30%-50%, avoiding stuttering, while high-performance devices fully utilize their hardware capabilities for optimal display effects, achieving a dynamic balance between resource utilization and display effect.

[0085] Furthermore, it employs a non-linear scaling algorithm combined with adaptive resolution adjustment based on graphic edge features, adaptive selection of GPU hardware texture compression format, and dynamic color space conversion between sRGB, DCI-P3, and Rec.2020. This replaces the linear scaling and fixed color space methods of existing technologies, ensuring consistent graphic clarity and color reproduction across different devices, improving frame rate stability by over 40%, and avoiding jagged edges and color deviations. Through real-time status monitoring and dynamic fine-tuning mechanisms, such as monitoring device status every 100ms and dynamically adjusting rendering strategies, the response latency is as low as ≤200ms, adapting to device load fluctuations and resolving the issue of sudden frame rate drops during load fluctuations. A display effect verification mechanism combining machine vision inspection and user experience feedback forms a closed-loop optimization loop, ensuring that the display effect continuously meets quality requirements.

[0086] It should be understood that, although Figure 1 The steps are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order in which these steps are executed; they can be performed in other orders. Figure 1At least some of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0087] In one embodiment, such as Figure 8 As shown, a cross-device adaptive graphics display control device 100 is also provided, including a device capability perception module 11, a graphics feature extraction module 13, an adaptation decision module 15, a real-time rendering module 17, and a display verification module 19. The device capability perception module 11 is used to determine the device capability information of the target display device by actively probing and passively reporting, acquiring the hardware parameters and software environment information of the target display device, and supplementing incomplete device capability information using a pre-built device fingerprint database; wherein, the device fingerprint database pre-stores parameter templates for mainstream display devices.

[0088] The graphics feature extraction module 13 uses a deep learning model to extract features from the raw graphics data of the input target display device, obtaining a standardized feature vector. The standardized feature vector includes graphics type, texture complexity, color space type, and dynamic feature information. The adaptation decision module 15 constructs a multi-dimensional adaptation decision model based on device capability information and the standardized feature vector, and outputs a rendering parameter adjustment scheme using a multi-objective optimization algorithm. The rendering parameter adjustment scheme includes resolution scaling ratio, texture compression level, color space conversion rules, and rendering level simplification strategy. The real-time rendering module 17 calls the graphics processing interface of the target display device to perform real-time rendering processing on the raw graphics data according to the rendering parameter adjustment scheme. The display verification module 19 collects the display output data of the rendered graphics and verifies the display effect. If the verification fails, it returns to the process of constructing the multi-dimensional adaptation decision model and readjusts the rendering parameter adjustment scheme.

[0089] The aforementioned cross-device adaptive graphics display control device achieves comprehensive acquisition of cross-device hardware parameters and software environment information through a device capability perception mechanism that combines active detection and passive reporting, and a device fingerprint database. It achieves automatic parsing of graphics features through a deep learning model, achieves dynamic balance between display effect and resource consumption through a multi-objective optimization algorithm, and achieves real-time adjustment of rendering strategy through real-time status monitoring and dynamic fine-tuning mechanism. This solves the problems of poor graphics display adaptability, unreasonable resource consumption, poor color consistency, and response lag in cross-device scenarios.

[0090] It is understood that the specific limitations of the cross-device adaptive graphics display control device 100 mentioned above can be understood by referring to the corresponding explanations and limitations of the various embodiments of the cross-device adaptive graphics display control method above, and will not be repeated here.

[0091] Each module in the aforementioned cross-device adaptive graphics display control device 100 can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of a device with data processing capabilities, or stored in software within the memory of the aforementioned device, so that the processor can call and execute the operations corresponding to each module. The aforementioned device can be, but is not limited to, various types of computer devices already existing in the art.

[0092] In one embodiment, a computer device is also provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the steps of the cross-device adaptation adaptive graphics display control method as described in the above embodiments.

[0093] In one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the cross-device adaptation adaptive graphics display control method as described in the above embodiments.

[0094] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), memory bus DRAM (RDRAM), and interface DRAM (DRDRAM), etc.

[0095] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0096] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of protection of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and all such modifications and improvements fall within the scope of protection of the present invention.

Claims

1. A cross-device adaptive graphics display control method, characterized in that, Including the following steps: By actively probing and passively reporting, the hardware parameters and software environment information of the target display device are obtained, and the device capability information of the target display device is determined. The incomplete device capability information is supplemented by a pre-built device fingerprint database. The device fingerprint database contains parameter templates of mainstream display devices. A deep learning model is used to extract features from the raw graphic data of the input target display device to obtain a standardized feature vector. The standardized feature vector includes graphic type, texture complexity, color space type, and dynamic feature information. Based on device capability information and standardized feature vectors, a multi-dimensional adaptation decision model is constructed and a multi-objective optimization algorithm is used to output a rendering parameter adjustment scheme. The rendering parameter adjustment scheme includes resolution scaling ratio, texture compression level, color space conversion rules, and rendering level simplification strategy. Based on the rendering parameter adjustment scheme, the graphics processing interface of the target display device is called to perform real-time rendering processing on the original graphics data. Collect the display output data of the rendered graphics and verify the display effect; if the verification fails, return to the process of building a multi-dimensional adaptation decision model and readjust the rendering parameter adjustment scheme.

2. The cross-device adaptive graphics display control method according to claim 1, characterized in that, The process of obtaining hardware parameters and software environment information of a target display device through active detection and passive reporting includes: Send a standardized query command to the target display device to obtain the hardware parameters of the target display device; the hardware parameters include display resolution, screen size, pixel density, memory capacity, GPU model and rendering capabilities; Obtain software environment information through the system interface of the target display device; software environment information includes operating system type and version, graphics API supported version, and display driver model; The process of supplementing incomplete equipment capability information using a pre-built equipment fingerprint database includes: When the device capability information obtained through active detection or passive reporting is incomplete, the missing information is supplemented by parameter templates in the device fingerprint database; the device fingerprint database matches basic parameters with the device model, and the parameter templates in the device fingerprint database are updated through the cloud server.

3. The cross-device adaptive graphics display control method according to claim 1, characterized in that, The steps for extracting features from the raw graphic data of the input target display device using a deep learning model to obtain standardized feature vectors include: The graphic type is identified based on the original graphic data; graphic types include 2D vector graphics, 2D bitmap graphics, 3D model graphics, and dynamic video frames; Texture complexity is quantified into multiple levels based on the quantity and density of detailed textures in the graphics; each level includes 1 to 10 levels indicating from simple to complex. Color space parameters are extracted from the original graphic data; the color space parameters include the color space type and RGB channel variance of the original graphic data. Dynamic features are extracted from dynamic video frames in the original graphic data; dynamic features include frame rate requirements and the amplitude of screen motion.

4. The cross-device adaptive graphics display control method according to any one of claims 1 to 3, characterized in that, The resolution scaling ratio is determined using a non-linear scaling algorithm based on the display resolution of the target display device and the original resolution of the original graphics data. The texture compression level is determined based on the compression format supported by the target display device's GPU and the texture complexity; the texture compression level includes 1 to 5 levels from no compression to the highest compression ratio, and matches the optimal compression format; Color space conversion rules include determining the conversion matrix between the color space of the original graphic data and the color space supported by the target display device; The rendering layer simplification strategy includes reducing the number of polygons in the model and reducing the complexity of lighting effects for 3D graphics based on the GPU rendering capabilities of the target display device.

5. The cross-device adaptive graphics display control method according to claim 4, characterized in that, Based on the rendering parameter adjustment scheme, the steps of calling the graphics processing interface of the target display device to perform real-time rendering processing of the original graphics data include: Based on the resolution scaling ratio, the graphics resolution is adjusted using a non-linear scaling algorithm through the scaling function of the graphics processing interface, and the resolution scaling ratio is adaptively adjusted based on the graphics edge features. Based on the texture compression level and compression format, the GPU hardware compression interface is invoked to compress the texture data; Color space conversion is performed through color matrix multiplication operations; the color space conversion includes dynamic switching between sRGB, DCI-P3 and Rec.2020; Rendering layers are simplified using a 3D model simplification algorithm; The device operation status of the target display device is monitored in real time at preset time intervals; when the load exceeds the preset upper limit threshold, the rendering complexity is dynamically reduced, and when the load is below the preset lower limit threshold, the display effect is improved. The load includes CPU load and GPU load.

6. The cross-device adaptive graphics display control method according to claim 4, characterized in that, The objective function of the multi-objective optimization algorithm is: Min F = α × (resource utilization rate) - β × (display performance score) Wherein, resource utilization rate = w1 × CPU utilization rate + w2 × GPU utilization rate + w3 × memory utilization rate, where w1 is the weighting coefficient of CPU utilization rate, w2 is the weighting coefficient of GPU utilization rate, and w3 is the weighting coefficient of memory utilization rate. Display performance score = w4 × Clarity score + w5 × Color reproduction score + w6 × Frame rate stability score, where w4 is the weighting coefficient for clarity, w5 is the weighting coefficient for color reproduction, w6 is the weighting coefficient for frame rate stability, α is the target weighting coefficient for resource utilization, and β is the target weighting coefficient for display performance score.

7. The cross-device adaptive graphics display control method according to claim 4, characterized in that, The steps for collecting and verifying the display output data of the rendered graphics include: The rendered graphics are captured using a screen capture device, and machine vision inspection metrics are calculated. These metrics include sharpness, color fidelity, and frame rate stability. Sharpness is calculated using peak signal-to-noise ratio (PSNR), with a preset threshold of at least 35 dB. Color fidelity is calculated by comparing the color deviation between the output image and the original image, with a preset threshold of at least 2. Frame rate stability is calculated by analyzing the actual displayed frame rate and determining the frame rate fluctuation, which is the difference between the maximum and minimum values ​​of the actual displayed frame rate. The preset frame rate stability threshold is at least 5%. The system obtains user ratings for the display effect through the configured user rating interface and calculates the average user rating. If the machine vision detection indicators do not meet the preset thresholds or the average user score is lower than the preset score threshold, the verification is deemed to have failed.

8. A cross-device adaptive graphic display control device, characterized in that, include: The device capability perception module is used to obtain the hardware parameters and software environment information of the target display device through active detection and passive reporting, and then determine the device capability information of the target display device. It uses a pre-built device fingerprint database to supplement incomplete device capability information; the device fingerprint database contains parameter templates of mainstream display devices. The graphics feature extraction module is used to extract features from the raw graphics data of the input target display device using a deep learning model to obtain a standardized feature vector. The standardized feature vector includes graphics type, texture complexity, color space type, and dynamic feature information. The adaptation decision module is used to construct a multi-dimensional adaptation decision model based on device capability information and standardized feature vectors, and to output a rendering parameter adjustment scheme using a multi-objective optimization algorithm. The rendering parameter adjustment scheme includes resolution scaling ratio, texture compression level, color space conversion rules, and rendering level simplification strategy. The real-time rendering module is used to adjust the scheme according to the rendering parameters and call the graphics processing interface of the target display device to perform real-time rendering processing of the original graphic data. The display verification module is used to collect the display output data of the rendered graphics and verify the display effect. If the verification fails, it returns to the process of building a multi-dimensional adaptation decision model and readjusts the rendering parameter adjustment scheme.

9. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the cross-device adaptive graphics display control method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the cross-device adaptive graphics display control method as described in any one of claims 1 to 7.

Images