A low-power processing video image quality enhancement method

Abstract

This invention relates to the field of video image processing technology and discloses a low-power video image quality enhancement method, including frame-by-frame video extraction, intelligent extraction of useful regions, local interpolation enhancement, and frame image synthesis steps. By focusing on key areas of user interest for targeted image quality optimization, it avoids the high power consumption problem of whole-frame processing, improving the clarity, contrast, and color saturation of core areas while reducing device computing power consumption and power consumption. This invention uses an image semantic segmentation algorithm and a key region determination model to automatically identify and extract useful areas of core user interest (such as people and main scenery), performing interpolation enhancement only on these local areas, focusing on improving clarity, contrast, and color saturation; non-critical areas retain their original image quality, avoiding ineffective enhancement, ensuring clear and prominent visual focus, and improving the overall viewing experience.

Description

Technical Field

[0001] This invention relates to the field of video image processing technology, specifically a low-power video image quality enhancement method. Background Technology

[0002] Video image enhancement technology is widely used in various terminal devices and image processing scenarios, but traditional methods have many technical limitations: First, they consume a lot of power and computing power. Most technologies adopt a full-frame image processing mode, performing contrast enhancement, sharpening, and color enhancement on all areas of each frame, which requires a lot of CPU resources, resulting in excessive power consumption. This is especially unsuitable for low-power scenarios such as mobile devices and embedded devices. Second, they have low processing efficiency. Full-frame processing requires calculation on massive amounts of pixel data, resulting in long processing times and difficulty in meeting the real-time processing requirements of video streams. Third, they lack targeted enhancement. They do not distinguish between key and non-key areas of interest to users, wasting computing power on meaningless background areas, and the enhancement effect on core areas is not prominent. Fourth, they have limited adaptability. They mostly use fixed interpolation algorithms and cannot flexibly adjust according to the image characteristics of different areas (such as blurriness and the degree of detail loss), resulting in inconsistent enhancement effects. Fifth, they lack a closed-loop optimization mechanism. After processing, the image quality effect is not verified, which can easily lead to over-enhancement or under-enhancement, making it difficult to guarantee stability.

[0003] While some related patented technologies already exist in the industry, significant shortcomings remain: Patent CN202210876543.9 (A video image quality enhancement method): It adopts a whole-frame image enhancement scheme, which consumes a lot of power and computing power and is not suitable for low-power devices; it does not distinguish between key areas and non-key areas, resulting in insufficient enhancement targeting and limited improvement in core areas; it only uses a single interpolation algorithm, which has poor adaptability and cannot cope with areas with different image features.

[0004] Patent CN202121987654.7 (Low-power image enhancement processing device): Although it adopts a local processing approach, the selection of key areas relies on manual presets, lacks intelligent recognition capabilities, and has low accuracy; it lacks image quality verification and closed-loop optimization mechanisms, resulting in insufficient stability of the enhancement effect; and it does not set up a cache management strategy, resulting in low processing efficiency.

[0005] Patent CN202320567890.1 (Real-time Video Quality Enhancement System): It supports real-time processing, but still focuses on whole-frame enhancement, resulting in poor power consumption control; the interpolation algorithm library is limited and cannot be dynamically adapted to regional features; the synthesized image is prone to splicing marks and unnatural transitions.

[0006] Patent CN202010987654.2 (Image Enhancement Method for Embedded Devices): Designed for low-power scenarios, but the accuracy of key area extraction is low, and areas are easily missed or misselected; there is no user-defined attention weight function, making it difficult to meet personalized needs; the smoothness of the video after reconstruction is easily affected, and there are differences from the original video.

[0007] The aforementioned existing technologies have not solved the core problems of "low-power processing - precise enhancement of key areas - high adaptability - efficient processing - stable effect", and are difficult to meet the needs of video image quality enhancement in low computing power and low power consumption scenarios. There is an urgent need for a targeted and optimized processing method. Summary of the Invention

[0008] The purpose of this invention is to provide a low-power video image quality enhancement method to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a low-power video image quality enhancement method, comprising the steps of video frame-by-frame extraction, intelligent extraction of useful regions, local interpolation enhancement, and frame image synthesis. By focusing on key regions of interest to the user for targeted image quality optimization, the high power consumption problem of whole-frame processing is avoided. While improving the clarity, contrast, and color saturation of core regions, the device's computing power consumption and power consumption are reduced.

[0010] Preferably, the specific process of the video frame-by-frame extraction step is as follows: The video parsing module establishes a data transmission connection with the frame extraction module. The video parsing module encodes and parses the input video file, extracts basic parameters such as the frame rate, resolution, and encoding format of the video stream, and transmits them to the frame extraction module. The frame extraction module extracts video images frame by frame according to the preset frame rate, synchronizes the original image data of each frame to the subsequent useful area extraction module, and retains a backup of the original frame file to ensure that the original image information is not lost in the subsequent synthesis process.

[0011] Preferably, the specific process of the intelligent extraction of the useful area is as follows: A two-way data interaction connection is established between the frame extraction module and the region identification and extraction module. The frame extraction module transmits the original image data of each frame to the region identification and extraction module. The region identification and extraction module has built-in image semantic segmentation algorithm and key region determination model. By analyzing the pixel features, texture information and user attention weight of each region in the image, it automatically identifies and accurately extracts the useful regions that the user is most concerned about (such as specific objects, main scenes, etc.). The extracted local region data and the non-key region data of the original image are marked and fed back to the frame extraction module for temporary storage, so as to achieve accurate positioning and efficient extraction of key regions.

[0012] Preferably, the local interpolation enhancement step is performed as follows: A data transmission connection is established between the region identification and extraction module and the interpolation enhancement module. The region identification and extraction module transmits the marked local useful region data to the interpolation enhancement module. The interpolation enhancement module uses an adaptive interpolation algorithm to supplement details for pixel loss in local areas. By improving the clarity, contrast and color saturation of the area, the image quality is enhanced. The entire processing is carried out only on the extracted local area. The computing power and power consumption required for processing are only about 20% of those for processing the entire frame, which significantly reduces the CPU load of the device.

[0013] Preferably, the frame image synthesis step is carried out as follows: The interpolation enhancement module establishes a data interaction connection with the image synthesis module. The interpolation enhancement module transmits the local area data after image quality enhancement to the image synthesis module. At the same time, the frame extraction module synchronizes the backup data of the non-critical area of ​​the original frame to the image synthesis module. The image synthesis module uses a pixel alignment algorithm to accurately fuse the enhanced local area with the non-critical area of ​​the original frame, restore the complete structure and spatial logic of the image, and generate single-frame optimized image data to ensure that the synthesized image has no splicing traces and a natural transition.

[0014] Preferably, the region identification and extraction module establishes an association connection with the user attention weight configuration module. The user attention weight configuration module allows users to customize the feature parameters of key objects of interest (such as color range, shape features, texture type, etc.) and transmits the configured weight parameters to the region identification and extraction module. The region identification and extraction module combines the user-defined parameters with the built-in recognition model to further improve the accuracy of useful region extraction and meet the image quality enhancement needs of different users for specific scenes or objects.

[0015] Preferably, the interpolation enhancement module has a built-in library of multiple interpolation algorithms, including bilinear interpolation, bicubic interpolation, and deep learning interpolation algorithms. The interpolation enhancement module establishes a data connection with the image feature analysis module. The image feature analysis module performs feature analysis on the extracted local area, such as blurriness and the degree of detail loss, and transmits the analysis results to the interpolation enhancement module. The interpolation enhancement module automatically selects the appropriate interpolation algorithm according to the regional features to achieve targeted detail supplementation and image quality optimization, ensuring that the enhancement effect matches the regional features.

[0016] Preferably, the image synthesis module establishes a data feedback connection with the image quality verification module. The image synthesis module transmits the synthesized single-frame image data to the image quality verification module. The image quality verification module compares the sharpness, contrast, and color saturation parameters of the enhanced local area with the corresponding area of ​​the original frame to determine whether the enhancement effect meets the standard. If it does not meet the standard, it sends a parameter adjustment instruction to the interpolation enhancement module. The interpolation enhancement module re-optimizes and processes the data before transmitting it back to the image synthesis module, forming a closed-loop process of "extraction-enhancement-synthesis-verification-optimization" to ensure the stability and effectiveness of image quality enhancement.

[0017] Preferably, in the video frame-by-frame extraction step, the frame extraction module establishes a data storage connection with the cache module. The frame extraction module temporarily stores the original image data of each frame to the cache module. The cache module adopts a hierarchical storage mechanism, prioritizing the caching of the currently processed frame and adjacent frame data to reduce the time spent on repeated data reading and improve overall processing efficiency. At the same time, the low-power management unit of the cache module dynamically adjusts the caching strategy according to the device operating status to further reduce device power consumption.

[0018] Preferably, the image synthesis module establishes a data transmission connection with the video reconstruction module. The image synthesis module transmits the optimized image data frame by frame to the video reconstruction module in a time sequence. The video reconstruction module reconstructs the continuously optimized single-frame images into a complete video file based on parameters such as the frame rate and encoding format of the original video. This ensures that the smoothness of the reconstructed video is consistent with the original video, while maintaining the image quality enhancement effect in the core area, thereby achieving an overall improvement in video image quality under the premise of low power consumption.

[0019] This invention provides a low-power video image quality enhancement method. It has the following beneficial effects: 1. This invention uses an image semantic segmentation algorithm and a key region determination model to automatically identify and extract useful regions (such as people and main scenery) that the user is most concerned about. Interpolation enhancement is performed only on these local regions to improve clarity, contrast and color saturation. Non-critical regions retain their original image quality, avoiding ineffective enhancement, ensuring that the visual focus is clear and prominent, and improving the overall viewing experience.

[0020] 2. This invention processes only the extracted local area, requiring only about 20% of the computing power and power consumption of processing the entire frame, significantly reducing the CPU load on the device; combined with a hierarchical caching mechanism and a low-power caching strategy, it further reduces data reading time and energy consumption, perfectly adapting to the usage needs of low-computing-power and low-power devices such as mobile phones and embedded monitoring.

[0021] 3. This invention uses an interpolation enhancement module with a built-in library of multiple interpolation algorithms. Combined with image feature analysis, it automatically selects the appropriate algorithm based on the blurriness and lack of detail in the local area to supplement details in a targeted manner. The image synthesis module uses a pixel alignment algorithm to accurately fuse the enhanced local area with the non-critical area of ​​the original frame, resulting in a natural transition without splicing marks, ensuring the integrity and harmony of the image.

[0022] 4. This invention supports users to customize the feature parameters (such as color and shape) of key objects of interest, accurately extracting target areas to meet the personalized enhancement needs of different scenarios; it establishes a closed-loop process of "extraction-enhancement-compositing-verification-optimization", and monitors the enhancement effect in real time through the image quality verification module. If the effect is not up to standard, the parameters are automatically adjusted and re-optimized to ensure the stability and effectiveness of image quality enhancement. Attached Figure Description

[0023] Figure 1 This is a flowchart of the method steps of the present invention; Figure 2 This is a flowchart illustrating the video frame-by-frame extraction logic of the present invention. Figure 3 This is a flowchart illustrating the intelligent extraction logic of the useful area in this invention. Figure 4 This is a flowchart of the local interpolation enhancement logic of the present invention; Figure 5 This is a flowchart illustrating the user-defined weight configuration logic of this invention. Figure 6 The flowchart below shows the selection logic of the adaptive interpolation algorithm of this invention. Figure 7 This is a flowchart of the closed-loop logic for image quality verification in this invention; Figure 8 This is a flowchart of the low-power management logic of the cache module of the present invention; Figure 9 This is a flowchart illustrating the video reconstruction logic of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.

[0026] A preferred embodiment of the low-power video image quality enhancement method provided by the present invention is as follows: Figure 1-9 As shown: A low-power video image quality enhancement method includes four core steps: frame-by-frame video extraction, intelligent extraction of useful regions, local interpolation enhancement, and frame image synthesis. These steps work together to achieve low-power, high-precision image quality enhancement. The specific process is as follows: Step 1: Frame-by-frame video extraction A data transmission connection is established between the video parsing module and the frame extraction module. The video parsing module first encodes and parses the input video file, extracting basic parameters such as frame rate (e.g., 24fps, 30fps), resolution (e.g., 1920×1080, 1280×720), and encoding format (e.g., H.264, H.265), and transmits them to the frame extraction module. The frame extraction module extracts video images frame by frame according to a preset frame rate, synchronizing the original image data of each frame to the subsequent useful area extraction module, while retaining a backup of the original frame file (including non-critical area data) to ensure that the original image information is not lost during subsequent synthesis. The frame extraction module establishes a data storage connection with the cache module, temporarily storing the extracted original image data of each frame in the cache module. The cache module adopts a hierarchical storage mechanism, prioritizing the caching of the currently processed frame and adjacent frame data to reduce the time spent on repeated data reading and improve overall processing efficiency. At the same time, the low-power management unit of the cache module dynamically adjusts the caching strategy according to the device's operating status (e.g., battery level, computing power utilization rate) (e.g., caching more frames when the battery is high and reducing the cache when the battery is low), further reducing device power consumption.

[0027] Step 2: Intelligent Extraction of Useful Areas A bidirectional data interaction connection is established between the frame extraction module and the region identification and extraction module. The frame extraction module transmits the original image data of each frame to the region identification and extraction module. The region identification and extraction module has a built-in image semantic segmentation algorithm and a key region determination model. By analyzing the pixel features (such as color and brightness), texture information (such as edges and detail density), and user attention weights of each region in the image, it automatically identifies and accurately extracts the useful regions (such as specific objects, main scenes, faces, etc.) that the user is most concerned about. The region identification and extraction module is also linked to the user attention weight configuration module. Users can use this module to customize the feature parameters (such as color range, shape features, texture type, etc.) of key objects of interest. The configured weight parameters are transmitted to the region identification and extraction module, and combined with the built-in recognition model, the accuracy of useful region extraction is further improved. After extraction, the region identification and extraction module marks the extracted local region data and the non-key region data of the original image respectively, and feeds them back to the frame extraction module for temporary storage, realizing accurate positioning and efficient extraction of key regions.

[0028] Step 3: Local Interpolation Enhancement Step A data transmission connection is established between the region identification and extraction module and the interpolation enhancement module. The region identification and extraction module transmits the marked local useful region data to the interpolation enhancement module. The interpolation enhancement module has a built-in library of multiple interpolation algorithms, including bilinear interpolation, bicubic interpolation, and deep learning interpolation algorithms. The interpolation enhancement module also establishes a data connection with the image feature analysis module. The image feature analysis module performs feature analysis on the extracted local region, such as blurriness and the degree of detail loss, and transmits the analysis results (e.g., high blurriness, severe detail loss) to the interpolation enhancement module. The interpolation enhancement module automatically selects the appropriate interpolation algorithm based on the region features (e.g., deep learning interpolation algorithm for high blurriness regions, bilinear interpolation algorithm for regions with slight detail loss), and supplements the details for pixel loss in the local region, thereby enhancing image quality by improving the clarity, contrast, and color saturation of the region. The entire processing is performed only on the extracted local region, and the computing power and power consumption required for processing are only about 20% of those for processing the entire frame, significantly reducing the CPU load on the device.

[0029] Step 4: Frame Image Synthesis A data interaction connection is established between the interpolation enhancement module and the image synthesis module. The interpolation enhancement module transmits the enhanced local area data to the image synthesis module, while the frame extraction module synchronizes the backed-up non-critical area data of the original frame to the image synthesis module. The image synthesis module uses a pixel alignment algorithm (such as feature point matching-based alignment) to accurately fuse the enhanced local area with the non-critical area of ​​the original frame, restoring the complete structure and spatial logic of the image and generating optimized single-frame image data to ensure that the synthesized image has no stitching artifacts and a natural transition. A data feedback connection is established between the image synthesis module and the image quality verification module. The image quality verification module judges the enhancement effect by comparing the sharpness, contrast, and color saturation parameters of the enhanced local area with the corresponding area of ​​the original frame. If the image quality is not up to standard (e.g., sharpness improvement ≥30%, contrast optimization ±15%), the image quality verification module sends parameter adjustment instructions to the interpolation enhancement module (e.g., increase sharpness intensity, adjust color saturation gain). The interpolation enhancement module re-optimizes and processes the image before transmitting it back to the image synthesis module, forming a closed-loop process of "extraction-enhancement-synthesis-verification-optimization". The image synthesis module establishes a data transmission connection with the video reconstruction module and transmits the optimized image data frame by frame to the video reconstruction module in a time sequence. The video reconstruction module reconstructs the continuously optimized single-frame images into a complete video file based on parameters such as the frame rate and encoding format of the original video, ensuring that the smoothness of the reconstructed video is consistent with the original video, while maintaining the image quality enhancement effect in the core area.

[0030] In summary, this invention achieves a balance between accurate enhancement of core areas and low power consumption by designing a process of video frame-by-frame extraction, intelligent extraction of useful regions, local interpolation enhancement, and frame image synthesis, combined with intelligent recognition, adaptive algorithms, closed-loop verification, and cache optimization. This improves processing efficiency and effect stability, and is suitable for low-computing-power scenarios.

[0031] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0032] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A low-power video image quality enhancement method, characterized in that, It includes steps such as video frame-by-frame extraction, intelligent extraction of useful areas, local interpolation enhancement, and frame image synthesis. By focusing on key areas of interest to users, it performs targeted image quality optimization, avoiding the high power consumption problem of whole-frame processing. While improving the clarity, contrast, and color saturation of core areas, it reduces the device's computing power consumption and power consumption.

2. The low-power video image quality enhancement method according to claim 1, characterized in that, The specific process of the video frame-by-frame extraction step is as follows: The video parsing module establishes a data transmission connection with the frame extraction module. The video parsing module encodes and parses the input video file, extracts basic parameters such as the frame rate, resolution, and encoding format of the video stream, and transmits them to the frame extraction module. The frame extraction module extracts video images frame by frame according to the preset frame rate, synchronizes the original image data of each frame to the subsequent useful area extraction module, and retains a backup of the original frame file to ensure that the original image information is not lost in the subsequent synthesis process.

3. The low-power video image quality enhancement method according to claim 1, characterized in that, The specific process of the intelligent extraction of useful regions is as follows: A two-way data interaction connection is established between the frame extraction module and the region identification and extraction module. The frame extraction module transmits the original image data of each frame to the region identification and extraction module. The region identification and extraction module has built-in image semantic segmentation algorithm and key region determination model. By analyzing the pixel features, texture information and user attention weight of each region in the image, it automatically identifies and accurately extracts the useful regions that the user is most concerned about (such as specific objects, main scenes, etc.). The extracted local region data and the non-key region data of the original image are marked and fed back to the frame extraction module for temporary storage, so as to achieve accurate positioning and efficient extraction of key regions.

4. The low-power video image quality enhancement method according to claim 1, characterized in that, The specific process of the local interpolation enhancement step is as follows: A data transmission connection is established between the region identification and extraction module and the interpolation enhancement module. The region identification and extraction module transmits the marked local useful region data to the interpolation enhancement module. The interpolation enhancement module uses an adaptive interpolation algorithm to supplement details for pixel loss in local areas. By improving the clarity, contrast and color saturation of the area, the image quality is enhanced. The entire processing is carried out only on the extracted local area. The computing power and power consumption required for processing are only about 20% of those for processing the entire frame, which significantly reduces the CPU load of the device.

5. The low-power video image quality enhancement method according to claim 1, characterized in that, The specific process of the frame image synthesis step is as follows: The interpolation enhancement module establishes a data interaction connection with the image synthesis module. The interpolation enhancement module transmits the local area data after image quality enhancement to the image synthesis module. At the same time, the frame extraction module synchronizes the backup data of the non-critical area of ​​the original frame to the image synthesis module. The image synthesis module uses a pixel alignment algorithm to accurately fuse the enhanced local area with the non-critical area of ​​the original frame, restore the complete structure and spatial logic of the image, and generate single-frame optimized image data to ensure that the synthesized image has no splicing traces and a natural transition.

6. The low-power video image quality enhancement method according to claim 1, characterized in that, The region identification and extraction module establishes a connection with the user attention weight configuration module. The user attention weight configuration module allows users to customize the feature parameters of key objects of interest (such as color range, shape features, texture type, etc.) and transmits the configured weight parameters to the region identification and extraction module. The region identification and extraction module combines the user-defined parameters with the built-in recognition model to further improve the accuracy of useful region extraction and meet the image quality enhancement needs of different users for specific scenes or objects.

7. The low-power video image quality enhancement method according to claim 1, characterized in that, The interpolation enhancement module has a built-in library of multiple interpolation algorithms, including bilinear interpolation, bicubic interpolation, and deep learning interpolation algorithms. The interpolation enhancement module establishes a data connection with the image feature analysis module. The image feature analysis module performs feature analysis on the extracted local area, such as blurriness and the degree of detail loss, and transmits the analysis results to the interpolation enhancement module. The interpolation enhancement module automatically selects the appropriate interpolation algorithm according to the regional features to achieve targeted detail supplementation and image quality optimization, ensuring that the enhancement effect matches the regional features.

8. The low-power video image quality enhancement method according to claim 1, characterized in that, The image synthesis module establishes a data feedback connection with the image quality verification module. The image synthesis module transmits the synthesized single-frame image data to the image quality verification module. The image quality verification module compares the sharpness, contrast, and color saturation parameters of the enhanced local area with the corresponding area of ​​the original frame to determine whether the enhancement effect meets the standard. If it does not meet the standard, it sends a parameter adjustment instruction to the interpolation enhancement module. The interpolation enhancement module re-optimizes and processes the data before transmitting it back to the image synthesis module, forming a closed-loop process of "extraction-enhancement-synthesis-verification-optimization" to ensure the stability and effectiveness of image quality enhancement.

9. The low-power video image quality enhancement method according to claim 1, characterized in that, In the video frame-by-frame extraction step, the frame extraction module establishes a data storage connection with the cache module. The frame extraction module temporarily stores the original image data of each frame to the cache module. The cache module adopts a hierarchical storage mechanism, prioritizing the caching of the currently processed frame and adjacent frame data to reduce the time spent on repeated data reading and improve overall processing efficiency. At the same time, the low-power management unit of the cache module dynamically adjusts the caching strategy according to the device operating status to further reduce device power consumption.

10. The low-power video image quality enhancement method according to claim 1, characterized in that, The image synthesis module establishes a data transmission connection with the video reconstruction module. The image synthesis module transmits the optimized image data frame by frame to the video reconstruction module in a time sequence. The video reconstruction module reconstructs the continuously optimized single-frame images into a complete video file based on parameters such as the frame rate and encoding format of the original video. This ensures that the smoothness of the reconstructed video is consistent with the original video, while maintaining the image quality enhancement effect in the core area, thus achieving an overall improvement in video image quality under the premise of low power consumption.

Images