Image enhancement method based on human eye attention perception mechanism

Abstract

The invention discloses an image enhancement method based on a human eye attention perception mechanism, and belongs to the technical field of image processing, and the method comprises the steps: obtaining the gaze focus distribution data of human eyes in an original image, and obtaining a visual attention distribution diagram based on the gaze focus distribution data; based on the visual attention distribution map, obtaining a distance weight between each pixel in the image and the fixation focus, and based on the distance weight between each pixel and the fixation focus, obtaining a regional human eye saliency grading matrix; based on the regional human eye saliency grading matrix and the visual attention distribution diagram, obtaining a layered processing regional map; and the original image is processed according to the region priority based on the hierarchical processing region map, and image enhancement based on a human eye attention perception mechanism is realized. According to the invention, image quality and computing resource consumption can be effectively balanced, targeted image enhancement is realized, and the visual experience of a user is improved.

Description

Technical Field

[0001] The present invention belongs to the field of image processing technology, and in particular relates to an image enhancement method based on the human eye attention perception mechanism. Background Art

[0002] Image enhancement technology, a key branch of visual information processing, plays an irreplaceable role in improving image quality and optimizing user experience. This is particularly crucial in scenarios such as artificial intelligence, surveillance systems, and virtual reality. By enhancing image details, it can significantly improve information readability and subsequent intelligent application effectiveness, leading to continued research in this area of ​​interest.

[0003] However, current mainstream image enhancement methods often suffer from a common shortcoming: a lack of in-depth consideration of user subjective perception. Many technologies tend to uniformly process the entire image, ignoring the differences in human visual attention across different regions. This approach can lead to an unbalanced allocation of resources, with important areas under-optimized and less important areas over-processed, thus compromising overall image performance and real-time performance.

[0004] Against this backdrop, studying how to incorporate human visual attention mechanisms to guide image enhancement has become a pressing challenge. The core challenge lies in accurately capturing the human eye's focus when observing an image and translating this focus into a basis for determining the visual saliency of an image region. Due to the dynamic nature and individual variability of human gaze, relying solely on static models makes it difficult to accurately locate the region of interest. This uncertainty in positioning further complicates the design of a reasonable intensity distribution for the enhancement process, leading to a discrepancy between the detailed presentation of key image areas and user needs.

[0005] Therefore, how to build a dynamic and intelligent enhancement processing framework based on the gaze behavior of the human eye to achieve high-precision optimization of the area near the gaze point and gradually adjust the enhancement intensity as the distance increases is the key problem solved by the present invention. Summary of the Invention

[0006] In order to solve the above technical problems, the present invention proposes an image enhancement method based on the human eye attention perception mechanism to solve the problems existing in the above-mentioned prior art.

[0007] To achieve the above object, the present invention provides an image enhancement method based on the human eye's attention perception mechanism, comprising:

[0008] Obtaining gaze focus distribution data of a human eye in an original image, and obtaining a visual attention distribution map based on the gaze focus distribution data;

[0009] Based on the visual attention distribution map, obtaining a distance weight between each pixel in the image and the focus of gaze, and obtaining a regional human eye saliency grading matrix based on the distance weight between each pixel and the focus of gaze;

[0010] Obtaining a hierarchical processing area map based on the regional human eye saliency grading matrix and the visual attention distribution map;

[0011] The original image is processed according to the regional priority based on the hierarchical processing area map to achieve image enhancement based on the human eye attention perception mechanism.

[0012] Optionally, the process of obtaining a visual attention distribution map based on the gaze focus distribution data includes:

[0013] An initial distribution map of visual attention is constructed based on the gaze focus distribution data; it is determined whether the gaze focus distribution data is missing or abnormal, and if so, a data interpolation method is used to fill it to obtain a complete focus distribution data set; feature extraction is performed on the complete focus distribution data set to obtain visual attention distribution features under different image scenes; if there is an area in the initial distribution map of visual attention where the concentration of the visual attention distribution features reaches a preset condition, weighted processing is performed on the area where the concentration reaches the preset condition to enhance the initial distribution map of visual attention and obtain a visual attention distribution map.

[0014] Optionally, the process of obtaining a distance weight between each pixel in the image and the gaze focus based on the visual attention distribution map, and obtaining a regional human eye saliency ranking matrix based on the distance weight between each pixel and the gaze focus includes:

[0015] Based on the visual attention distribution map, the position data of each pixel in the image is obtained, and the spatial distance between each pixel and the focus data is determined based on the position data to obtain a distance weight distribution; based on a preset weight threshold and the distance weight distribution, a regional human eye saliency division result is obtained; based on the regional human eye saliency division result, the area that meets the human eye saliency requirement is extracted and the features are enhanced to obtain an enhanced regional feature distribution; through the enhanced regional feature distribution, the matching degree between the spatial distance of each area and the focus data is calculated, the priority level of the matching degree clustered area is adjusted, and the priority distribution is updated; according to the updated priority distribution, structured data of a hierarchical matrix is ​​generated and the human eye saliency ranking of each area in the overall distribution is judged to obtain a regional human eye saliency hierarchical matrix.

[0016] Optionally, the process of obtaining a hierarchical processing region map based on the regional human eye saliency grading matrix and the visual attention distribution map includes:

[0017] The regional human eye saliency grading matrix is ​​integrated with the visual attention distribution map to obtain a regional human eye saliency distribution map; the regional human eye saliency distribution map is graded according to a preset human eye saliency threshold to obtain a regional classification result; based on the regional classification result, the spatial distribution of each classified area is mapped to obtain a regional structural distribution and converted into a visual distribution map to obtain an initial regional map; association rules between resource allocation and regional human eye saliency are constructed, resources are allocated based on the association rules, and the allocation plan is integrated into the initial regional map to obtain a hierarchically processed regional map.

[0018] Optionally, the process of processing the original image according to the regional priority based on the hierarchical processing region map includes:

[0019] Based on the hierarchical processing area map, areas whose priorities exceed a threshold are obtained and detail enhancement processing is performed to obtain first image data; based on the first image data, areas whose priorities do not exceed a threshold are progressively smoothed to obtain second image data; difference values ​​between areas in the second image data are obtained, and based on the difference values, it is determined whether boundary transition enhancement processing is to be performed, and if boundary transition enhancement processing is to be performed, third image data is obtained; based on the third image data, an optimized demand distribution map is obtained, and local adaptive enhancement is performed based on the optimized demand distribution map to achieve image enhancement based on the human eye's attention perception mechanism.

[0020] Optionally, the process of obtaining the first image data includes:

[0021] Based on the hierarchical processing area map, the area whose priority exceeds the threshold is segmented to obtain a high-priority area; the original image data corresponding to the high-priority area is obtained in the original image, and the detail feature distribution of the original image data is obtained; based on the detail feature distribution, the original image data is enhanced using a histogram equalization method until the pixel distribution uniformity meets the preset conditions; the comparison data between the enhanced original image data and other areas in the original image is obtained, and if the difference in the comparison data is higher than the preset threshold, the optimization is continued to finally obtain the first image data.

[0022] Optionally, the process of obtaining the second image data includes:

[0023] Based on the hierarchical processing area map and the first image data, the area whose priority does not exceed the threshold is segmented to obtain a low-priority area; based on the human eye significance of each area in the hierarchical processing area map, different smoothing intensities are set for the low-priority area, and progressive smoothing processing and brightness compensation are performed until the uniformity of the brightness distribution meets the requirements; wherein, the resource occupancy information of the current task is obtained, and the hierarchical processing area map is updated according to the resource occupancy information of the current task.

[0024] Optionally, the process of obtaining the third image data includes:

[0025] Obtain difference distribution feature information of the second image data, locate the boundary area based on the difference distribution feature information and the edge detection algorithm to obtain boundary area distribution information; perform boundary transition enhancement based on the boundary area distribution information to obtain third image data.

[0026] Optionally, the texture distribution and color distribution of the third image data are scanned and detected to obtain and mark areas with insufficient details, thereby obtaining an optimization requirement distribution map.

[0027] Optionally, the process of performing local adaptive enhancement based on the optimized demand distribution map includes:

[0028] Based on the hierarchical processing area map, the priority of each area in the optimization demand distribution map is obtained, the optimization demand distribution map is layered according to the priority of each area, the layered area data is obtained, and the pixel enhancement of the layered area data is performed to achieve image enhancement based on the human eye attention perception mechanism.

[0029] Compared with the prior art, the present invention has the following advantages and technical effects:

[0030] The present invention obtains the distribution of user gaze focus through a pre-established eye tracking database, forms a visual attention distribution map, and calculates a regional human eye saliency grading matrix. According to the grading matrix, the present invention performs detail enhancement on high-priority areas, smoothes low-priority areas, and performs transition enhancement on boundary areas to maintain the overall consistency of the image. Combined with the user perception feedback database, the present invention also performs local adaptive enhancement on areas with insufficient detail, and finally fine-tunes the global image to obtain the final enhanced image. The present invention can effectively balance image quality and computing resource consumption, achieve targeted image enhancement, and enhance the user's visual experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0032] Figure 1 Schematic diagram of a method flow in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0034] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0035] Example 1

[0036] like Figure 1 As shown, this embodiment provides an image enhancement method based on the human eye attention perception mechanism, including:

[0037] Obtaining gaze focus distribution data of a human eye in an original image, and obtaining a visual attention distribution map based on the gaze focus distribution data;

[0038] As a specific implementation, the process of obtaining a visual attention distribution map based on gaze focus distribution data includes:

[0039] An initial distribution map of visual attention is constructed based on the gaze focus distribution data; it is determined whether the gaze focus distribution data is missing or abnormal, and if so, a data interpolation method is used to fill it to obtain a complete focus distribution data set; feature extraction is performed on the complete focus distribution data set to obtain visual attention distribution features under different image scenes; if there is an area in the initial distribution map of visual attention where the concentration of visual attention distribution features reaches a preset condition, weighted processing is performed on the area where the concentration reaches the preset condition to enhance the initial distribution map of visual attention and obtain a visual attention distribution map.

[0040] For example, when studying the distribution of users' visual attention across image scenes, a pre-built eye-tracking database can be used to obtain data on users' gaze focus in different scenarios. Suppose we analyze a set of indoor scene images. The database records the gaze distribution of 100 users on each image, but some areas are missing data due to users not paying attention. In this case, data interpolation methods can be used, such as filling in missing data based on the average of neighboring points, to ensure a complete focus distribution dataset. This method can effectively reduce the impact of incomplete data on subsequent analysis.

[0041] For example, during the feature extraction phase, a convolutional neural network model is applied to the complete focus distribution dataset. For example, given an image of an indoor scene, the model uses multiple layers of convolution to extract features of areas where users' attention is concentrated. For example, a certain decorative area in the image has an 80% attention density, far higher than the 20% in other areas. This significant concentration of features indicates that the area is more attractive to users. For example, weighted processing can be applied to these concentrated areas, such as increasing the attention weight of the area by 1.5 times, to generate an enhanced attention distribution map, thereby highlighting the visual salience of key areas.

[0042] Based on the visual attention distribution map, the distance weight between each pixel in the image and the focus of attention is obtained, and based on the distance weight between each pixel and the focus of attention, the regional human eye saliency classification matrix is ​​obtained;

[0043] As a specific implementation, the process of obtaining a distance weight between each pixel in the image and the focus of attention based on the visual attention distribution map, and obtaining a regional human eye saliency ranking matrix based on the distance weight between each pixel and the focus of attention includes:

[0044] Based on the visual attention distribution map, the position data of each pixel in the image is obtained, and the spatial distance between each pixel and the focus data is determined based on the position data to obtain the distance weight distribution; based on the preset weight threshold and the distance weight distribution, the regional human eye saliency division result is obtained; based on the regional human eye saliency division result, the area that meets the human eye saliency requirement is extracted and the features are enhanced to obtain the enhanced regional feature distribution; through the enhanced regional feature distribution, the matching degree between the spatial distance of each area and the focus data is calculated, the priority level of the matching degree clustered area is adjusted, and the priority distribution is updated; according to the updated priority distribution, the structured data of the hierarchical matrix is ​​generated and the human eye saliency ranking of each area in the overall distribution is judged to obtain the regional human eye saliency hierarchical matrix.

[0045] For example, when processing data related to visual attention distribution maps, we can first extract information about each pixel's position from the image and analyze its spatial relationship with the focus data. Assuming an image with a resolution of 1920x1080 and the focus data concentrated in the center of the image, we can initially construct a distance weight distribution map by calculating the distance from each pixel to the central focus. Pixels closer to the center have higher weights; for example, the center might have a weight of 0.9, while the edge might have a weight of only 0.2. This approach helps quickly distinguish the correlations between different regions in the image.

[0046] In one possible implementation, a threshold, such as 0.5, can be set for the initial distance weight distribution. Pixels with weights below this threshold are classified as low-visual-saliency regions. For example, pixel weights in edge regions are often below 0.5 and are therefore classified as low-visual-saliency regions, while central regions are classified as high-visual-saliency regions. This division provides a basic regional distribution map for subsequent analysis.

[0047] To enhance features in areas of high visual saliency, a convolutional neural network model can be used to perform in-depth analysis of the texture and color information in these areas. For example, if the central region contains a prominent target object, its feature distribution will be more prominent after processing by the model, enhancing the region's recognition within the overall image. This enhancement helps to more accurately identify key areas.

[0048] In one possible implementation, when calculating the match between regional feature distribution and focal data, the concentration of pixels within regions of high visual saliency can be analyzed. If the match in the central region reaches over 80%, while the matching in the surrounding regions is only 30%, the central region is prioritized to increase its visual saliency ranking. This adjustment ensures the targeted nature of subsequent analysis.

[0049] When generating a grading matrix, the image can be divided into multiple subregions, each of which is assigned a different grading value based on the priority distribution. For example, the central region is assigned grading 5, while the peripheral regions are assigned grading 1. This structured data facilitates subsequent processing and analysis. Next, using data mapping techniques, the matrix information is merged with the original distribution map to generate the final grading result.

[0050] Based on the regional human eye saliency classification matrix and visual attention distribution map, a hierarchical processing area map is obtained;

[0051] As a specific implementation, the process of obtaining a hierarchical processing region map based on the regional human eye saliency classification matrix and the visual attention distribution map includes:

[0052] The regional human eye saliency classification matrix is ​​integrated with the visual attention distribution map to obtain the regional human eye saliency distribution map; the regional human eye saliency map is graded according to the preset human eye saliency threshold to obtain the regional classification result; based on the regional classification result, the spatial distribution of each classification area is mapped to obtain the regional structure distribution and converted into a visual distribution map to obtain the initial regional map; the association rules between resource allocation and regional human eye saliency are constructed, and resources are allocated based on the association rules. The allocation plan is integrated into the initial regional map to obtain a hierarchical processing regional map.

[0053] Specifically, if the distribution data of a certain area is higher than the threshold, it is classified as high priority, and the hierarchical regional classification results are determined. Based on the hierarchical regional classification results, a hierarchical processing framework is constructed according to the business logic of hierarchical processing. The spatial distribution of high-priority and low-priority corresponding areas is mapped to obtain the regional structural distribution of hierarchical processing.

[0054] Specifically, when applying a preset threshold to divide high- and low-priority applications, a weight threshold of 0.7 can be set. Areas above this threshold are classified as high-priority. For example, if the center of an image has a weight of 0.85, it is classified as high-priority, while the edge of the image has a weight of 0.3, making it a low-priority area. This division method facilitates the rationality of subsequent resource allocation.

[0055] Specifically, the spatial distribution of high-priority and low-priority areas is mapped into a hierarchical structure. Assuming the high-priority area in the center of the image occupies 30% of the total area, it can be layered and stored with other areas using spatial coordinates to form a clear regional structure. This framework facilitates rapid identification of key areas during subsequent processing.

[0056] When building a regional map, spatial mapping technology can be used to transform the structural distribution into a visual distribution map. For example, using color coding, high-priority areas are marked in red, and low-priority areas are marked in blue, forming an initial regional map. This visualization method facilitates an intuitive understanding of regional distribution.

[0057] Given the correlation between resource allocation and priority, if the central area is classified as a high-priority area, processing resources are allocated first. For example, if there are 100 total resources, 70 units are allocated to the high-priority area and 30 units to the low-priority area. This prioritization ensures that critical areas receive adequate processing.

[0058] Based on the hierarchical processing area map, the original image is processed according to the area priority to achieve image enhancement based on the human eye's attention perception mechanism.

[0059] As a specific implementation, the process of processing the original image according to the regional priority based on the hierarchical processing region map includes:

[0060] Based on the hierarchical processing area map, areas with priorities exceeding a threshold are obtained and detail enhancement processing is performed to obtain first image data; based on the first image data, areas with priorities not exceeding a threshold are progressively smoothed to obtain second image data; difference values ​​between areas in the second image data are obtained, and based on the difference values, it is determined whether boundary transition enhancement processing is to be performed. If boundary transition enhancement processing is to be performed, third image data is obtained; based on the third image data, an optimized demand distribution map is obtained, and local adaptive enhancement is performed based on the optimized demand distribution map to achieve image enhancement based on the human eye's attention perception mechanism.

[0061] As a specific implementation, the process of obtaining the first image data includes:

[0062] Based on the hierarchical processing area map, the area with a priority exceeding a threshold is segmented into regional boundaries to obtain a high-priority area; the original image data corresponding to the high-priority area is obtained in the original image, and the detail feature distribution of the original image data is obtained; based on the detail feature distribution, the original image data is enhanced using a histogram equalization method until the pixel distribution uniformity meets the preset conditions; the comparison data between the enhanced original image data and other areas in the original image is obtained, and if the difference in the comparison data is higher than the preset threshold, the optimization is continued to finally obtain the first image data.

[0063] Specifically, by processing the regional map in layers, spatial analysis tools are used to finely segment the regional boundaries based on the distribution characteristics of the high-priority areas, thereby obtaining the precise range distribution of the high-priority areas. Based on the precise range distribution of the high-priority areas, and in accordance with the business needs of detail enhancement, the original image data within the area is obtained, and a preset filtering tool is applied to preliminarily extract the image details to determine the distribution of detail features. Based on the distribution of detail features, and in accordance with the specific implementation of the enhancement operation, the histogram equalization method in the image enhancement algorithm is used to adjust the brightness and contrast of the detail features to obtain the enhanced initial image data. Based on the enhanced initial image data, and in accordance with the business logic of intensity adjustment, the pixel distribution characteristics of the image data are obtained. If the uniformity of the pixel distribution is lower than a preset threshold, intensity compensation is performed on the local area to determine the image balance state after intensity adjustment. Based on the image balance state after intensity adjustment, and in accordance with the integration needs of regional division and processing priority, comparative data between the high-priority area and other areas is obtained. If the difference in the comparative data is higher than a preset threshold, secondary detail optimization is performed on the high-priority area to ultimately obtain the first image data.

[0064] For example, when acquiring raw image data and extracting details for precise range distribution in high-priority areas, filtering tools can be used to perform preliminary processing on texture and edge information within the image. Assuming the image data contains roads and vehicles, filtering tools can highlight vehicle outlines and road markings, creating a detailed feature distribution map that provides foundational data for subsequent enhancements.

[0065] For example, when performing image enhancement based on the distribution of detailed features, histogram equalization can adjust brightness and contrast. For example, if some areas in the original image are blurred due to insufficient lighting, this method can increase the brightness from an average of 50 to 80, ensuring that details are more clearly visible. This adjustment significantly improves the accuracy of subsequent region recognition.

[0066] Intensity adjustment logic applies intensity compensation to local areas if the pixel distribution uniformity falls below a preset threshold of 0.6. For example, if the pixel values ​​in a certain area are concentrated in a low-brightness range, compensation can increase the overall uniformity to above 0.75, improving image balance and providing more reliable data for regional comparison.

[0067] As a specific implementation, the process of obtaining the second image data includes:

[0068] Based on the hierarchical processing area map and the first image data, the area whose priority does not exceed the threshold is segmented to obtain a low-priority area; based on the human eye saliency of each area in the hierarchical processing area map, different smoothing intensities are set for the low-priority area, and progressive smoothing processing and brightness compensation are performed until the brightness distribution uniformity meets the requirements; wherein, the resource occupancy information of the current task is obtained, and the hierarchical processing area map is updated according to the resource occupancy information of the current task.

[0069] Specifically, the first image data is used to segment the region whose priority does not exceed the threshold, and obtain a low-priority region. For the original image data of the low-priority region, a progressive smoothing technique is applied to perform a preliminary adjustment on the pixel distribution in the region to obtain smoothed intermediate image data. Through the smoothed intermediate image data, the brightness distribution characteristics of the image data are obtained based on the overall balance requirements of the visual presentation. If the uniformity of the brightness distribution is lower than the preset threshold, the local area is brightness compensated to determine the compensated image data. Based on the compensated image data, the allocation of computing resources is optimized to obtain the resource occupancy of the current processing task. If the resource occupancy exceeds the preset limit, the processing operations are prioritized to obtain an adjusted resource allocation plan. For the processing operations in the low-priority region, a dynamic adjustment tool is used to perform hierarchical control on the intensity of the smoothing operation to obtain the optimized second image data.

[0070] Applying progressive smoothing to image data in low-priority areas can be understood as adjusting the distribution by gradually reducing the degree of abrupt changes between pixels. Assuming the original image data contains noise points, the smoothing technique averages the values ​​in 3x3 pixel units, gradually reducing the abrupt changes and generating intermediate image data. This process effectively improves visual smoothness.

[0071] During the brightness distribution analysis, if uniformity falls below a preset threshold, such as 0.6, local brightness compensation is required. For example, if the brightness values ​​in a certain area are concentrated between 20 and 50, compensation can be performed by increasing the dark areas to 60 to ensure a more balanced overall brightness distribution. This adjustment helps improve image visualization quality.

[0072] Computing resource allocation can be optimized by monitoring the resource usage of the current task. If a task occupies more than 80% of memory, priority sorting will postpone processing tasks in lower-priority areas, freeing up resources for higher-priority tasks. This dynamic allocation ensures system stability.

[0073] In the Dynamic Adjustment tool, the smoothing intensity is controlled in layers, allowing different smoothing parameters to be set based on the visual saliency of the region. For example, if the smoothing intensity is set to 0.3 for the edge region and 0.5 for the central transition region, the resulting second image data will exhibit a more natural transition effect. This layered control optimizes resource efficiency.

[0074] As a specific implementation, the process of obtaining the third image data includes:

[0075] Obtain difference distribution feature information of the second image data, locate the boundary area based on the difference distribution feature information and the edge detection algorithm to obtain boundary area distribution information; perform boundary transition enhancement based on the boundary area distribution information to obtain third image data.

[0076] Specifically, the second image data is analyzed using a preset analysis tool to scan the difference distribution for regional differences, obtain characteristic information about the difference distribution, and determine a preliminary difference range. Based on this preliminary difference range, an edge detection algorithm is applied to locate the boundary region based on its characteristics, obtaining precise distribution information about the boundary region and generating located boundary data. Based on this located boundary data, a progressive enhancement algorithm is used to adjust the pixel distribution in the boundary region for transitional enhancement. Enhanced intermediate data is obtained, and the adjusted image features are determined to generate the third image data.

[0077] As a specific implementation manner, the texture distribution and color distribution of the third image data are scanned and detected, and areas with insufficient details are obtained and marked to obtain an optimization requirement distribution map.

[0078] As a specific implementation method, the process of performing local adaptive enhancement based on the optimized demand distribution map includes:

[0079] Based on the hierarchical processing of the regional map, the priority of each area in the optimization demand distribution map is obtained, the optimization demand distribution map is layered according to the priority of each area, the layered regional data is obtained, and the pixels of the layered regional data are enhanced to achieve image enhancement based on the human eye's attention perception mechanism.

[0080] Specifically, for areas with insufficient detail in the optimization requirement distribution map, a preset image analysis tool is used to scan the marked areas, obtain feature distribution information for the deficient areas, and determine the priority range for local enhancement based on the layered processing area map. Based on the priority range for local enhancement, the marked areas are layered using image adjustment tools to determine the desired adaptive adjustment requirements. Layered regional data is obtained and the initial enhancement direction is determined. If the feature distribution of the layered regional data does not meet a preset threshold, weights are adjusted for the specific marked areas, and the adjusted weight distribution information is obtained to determine the specific adaptive enhancement plan. Based on the adjusted weight distribution information, a convolutional neural network in a deep learning model is used to extract detail features from the marked areas to optimize image content. Extracted detail feature data is obtained and the applicable scope of enhancement is determined. If the detail feature data deviates from the preset clarity standard, the marked areas are subjected to local pixel-level enhancement to obtain enhanced regional image information and determine the final adjusted detail content. The adjusted detail content is then fused with the overall image content to obtain fused image data, which serves as the final image output, achieving image enhancement based on the human eye's attention perception mechanism.

[0081] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. An image enhancement method based on the human eye's attention perception mechanism, characterized in that: The following steps are involved: Obtaining gaze focus distribution data of a human eye in an original image, and obtaining a visual attention distribution map based on the gaze focus distribution data; Based on the visual attention distribution map, obtaining a distance weight between each pixel in the image and the focus of gaze, and obtaining a regional human eye saliency grading matrix based on the distance weight between each pixel and the focus of gaze; Obtaining a hierarchical processing area map based on the regional human eye saliency grading matrix and the visual attention distribution map; The original image is processed according to the regional priority based on the hierarchical processing area map to achieve image enhancement based on the human eye attention perception mechanism.

2. The image enhancement method based on the human eye attention perception mechanism according to claim 1, characterized in that: The process of obtaining a visual attention distribution map based on gaze focus distribution data includes: An initial distribution map of visual attention is constructed based on the gaze focus distribution data; it is determined whether the gaze focus distribution data is missing or abnormal, and if so, a data interpolation method is used to fill it to obtain a complete focus distribution data set; feature extraction is performed on the complete focus distribution data set to obtain visual attention distribution features under different image scenes; if there is an area in the initial distribution map of visual attention where the concentration of the visual attention distribution features reaches a preset condition, weighted processing is performed on the area where the concentration reaches the preset condition to enhance the initial distribution map of visual attention and obtain a visual attention distribution map.

3. The image enhancement method based on the human eye attention perception mechanism according to claim 1, characterized in that: The process of obtaining a distance weight between each pixel in the image and the focus of gaze based on the visual attention distribution map, and obtaining a regional human eye saliency grading matrix based on the distance weight between each pixel and the focus of gaze includes: Based on the visual attention distribution map, the position data of each pixel in the image is obtained, and the spatial distance between each pixel and the focus data is determined based on the position data to obtain a distance weight distribution; based on a preset weight threshold and the distance weight distribution, a regional human eye saliency division result is obtained; based on the regional human eye saliency division result, the area that meets the human eye saliency requirement is extracted and the features are enhanced to obtain an enhanced regional feature distribution; through the enhanced regional feature distribution, the matching degree between the spatial distance of each area and the focus data is calculated, the priority level of the matching degree clustered area is adjusted, and the priority distribution is updated; according to the updated priority distribution, structured data of a hierarchical matrix is ​​generated and the human eye saliency ranking of each area in the overall distribution is judged to obtain a regional human eye saliency hierarchical matrix.

4. The image enhancement method based on the human eye attention perception mechanism according to claim 1, characterized in that: The process of obtaining a hierarchical processing area map based on the regional human eye saliency grading matrix and the visual attention distribution map includes: The regional human eye saliency grading matrix is ​​integrated with the visual attention distribution map to obtain a regional human eye saliency distribution map; the regional human eye saliency distribution map is graded according to a preset human eye saliency threshold to obtain a regional classification result; based on the regional classification result, the spatial distribution of each classified area is mapped to obtain a regional structural distribution and converted into a visual distribution map to obtain an initial regional map; association rules between resource allocation and regional human eye saliency are constructed, resources are allocated based on the association rules, and the allocation plan is integrated into the initial regional map to obtain a hierarchically processed regional map.

5. The image enhancement method based on the human eye attention perception mechanism according to claim 4, characterized in that: The process of processing the original image according to the regional priority based on the hierarchical processing area map includes: Based on the hierarchical processing area map, areas whose priorities exceed a threshold are obtained and detail enhancement processing is performed to obtain first image data; based on the first image data, areas whose priorities do not exceed a threshold are progressively smoothed to obtain second image data; difference values ​​between areas in the second image data are obtained, and based on the difference values, it is determined whether boundary transition enhancement processing is to be performed, and if boundary transition enhancement processing is to be performed, third image data is obtained; based on the third image data, an optimized demand distribution map is obtained, and local adaptive enhancement is performed based on the optimized demand distribution map to achieve image enhancement based on the human eye's attention perception mechanism.

6. The image enhancement method based on the human eye attention perception mechanism according to claim 5, characterized in that: The process of obtaining the first image data includes: Based on the hierarchical processing area map, the area whose priority exceeds the threshold is segmented to obtain a high-priority area; the original image data corresponding to the high-priority area is obtained in the original image, and the detail feature distribution of the original image data is obtained; based on the detail feature distribution, the original image data is enhanced using a histogram equalization method until the pixel distribution uniformity meets the preset conditions; the comparison data between the enhanced original image data and other areas in the original image is obtained, and if the difference in the comparison data is higher than the preset threshold, the optimization is continued to finally obtain the first image data.

7. The image enhancement method based on the human eye attention perception mechanism according to claim 5, characterized in that: The process of obtaining the second image data includes: Based on the hierarchical processing area map and the first image data, the area whose priority does not exceed the threshold is segmented to obtain a low-priority area; based on the human eye significance of each area in the hierarchical processing area map, different smoothing intensities are set for the low-priority area, and progressive smoothing processing and brightness compensation are performed until the uniformity of the brightness distribution meets the requirements; wherein, the resource occupancy information of the current task is obtained, and the hierarchical processing area map is updated according to the resource occupancy information of the current task.

8. The image enhancement method based on the human eye attention perception mechanism according to claim 5, characterized in that: The process of obtaining the third image data includes: Obtain difference distribution feature information of the second image data, locate the boundary area based on the difference distribution feature information and the edge detection algorithm to obtain boundary area distribution information; perform boundary transition enhancement based on the boundary area distribution information to obtain third image data.

9. The image enhancement method based on the human eye attention perception mechanism according to claim 5, characterized in that: The texture distribution and color distribution of the third image data are scanned and detected to obtain and mark areas with insufficient details, thereby obtaining an optimization requirement distribution map.

10. The image enhancement method based on human eye attention perception mechanism according to claim 5, characterized in that: The process of performing local adaptive enhancement based on the optimized demand distribution graph includes: Based on the hierarchical processing area map, the priority of each area in the optimization demand distribution map is obtained, the optimization demand distribution map is layered according to the priority of each area, the layered area data is obtained, and the pixel enhancement of the layered area data is performed to achieve image enhancement based on the human eye attention perception mechanism.

Images