Light-load video coding method and system based on image texture and gradient features

Abstract

The invention discloses a light-load video coding method and system based on image texture and gradient characteristics, and relates to the technical field of image management. A picture before compression processing is segmented, the image texture characteristics of each segmented unit region are extracted, the gradient of the texture characteristics among different unit regions is obtained, and the gradient of the texture characteristics of each unit region is obtained; and taking the maximum gradient of the unit region and the adjacent region as an association feature between the unit region and the adjacent region, performing feature extraction on the association feature in the picture through a convolutional network to obtain a feature vector of each unit region in the picture, performing texture feature evaluation on the compressed picture after the picture is compressed, and obtaining a feature vector of each unit region in the picture. And comparing the characteristic vector with the characteristic vector obtained before compression, screening out a unit region with abnormal gradient for sharpening compensation, and controlling the file volume while improving the file quality of the compressed picture.

Description

Technical Field

[0001] The present invention relates to the technical field of image management, and specifically to a light-load video coding method and system based on image texture and gradient features. Background Art

[0002] With the increasing applications of image recording devices such as cameras and video cameras in industrial and living scenarios, and the continuous increase in the volume of a single image, new challenges are posed to image storage. Therefore, in existing common solutions, it is necessary to compress the original image recording before storage to reduce the storage space of image information.

[0003] However, after image compression, some image detail information is discarded, resulting in the problem of image blurring and detail loss. In application scenarios with extremely high requirements for image details such as industrial inspection, it is more likely to cause the omission of key information, thereby affecting production decision-making and quality assessment. For images with blurred details, they usually need to be discarded or undergo detail enhancement. In the prior art, the methods for detail enhancement need to process the entire image, and all image coding information of the image needs to be processed in the computer system, causing an operating burden on the image processing system. Summary of the Invention

[0004] The purpose of the present invention is to provide a light-load video coding method and system based on image texture and gradient features to solve the problems proposed in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A light-load video coding method based on image texture and gradient features, the method comprising: Step S100: Segment the picture before compression processing, and extract the image texture features of each unit area after segmentation; Step S200: Obtain the gradient of the texture features between different unit areas, and use the maximum gradient between the unit area and the adjacent area as the association feature between the unit area and the adjacent area; Step S300: Extract the association features in the picture through a convolutional network to obtain the feature vectors of each unit area in the picture; Step S400: After the picture is compressed, perform texture feature evaluation on the compressed picture, compare it with the feature vectors obtained before compression, and screen out the unit areas with abnormal gradients for sharpening compensation.

[0006] Further, step S100 includes: Step S101: Obtain the picture before compression processing, denote the picture as the first target picture, segment the first target picture into N unit areas, and extract the gray feature value and texture feature value of each unit area; Step S102: Compose the gray-scale feature values of each unit area into a gray-scale feature vector, and compose the texture feature values of each unit area into a texture feature vector; Step S103: Concatenate the gray-scale feature vector and the texture feature vector of each unit area into a region feature vector corresponding to each unit area; Step S104: Calculate the gray-scale feature values of each unit area, where the gray-scale feature values include calculating the contrast of the gray-scale image, the uniformity of the gray-scale distribution, the similarity of the gray-scale distribution, and the correlation of the gray-scale values; Among them, the uniformity of the gray-scale distribution is calculated by the inverse difference moment of the GLCM matrix. The larger the inverse difference moment, the more consistent the gray-scale levels in the image, the more uniform the texture. The similarity of the gray-scale distribution is calculated by the homogeneity of the GLCM matrix, and the correlation of the gray-scale values is calculated by the correlation of the GLCM matrix.

[0007] Further, step S200 includes: Step S201: Obtain the centroid of the unit area. When the distance between the centroid of a certain unit area and the centroid of another unit area is less than the distance threshold, record the other unit area as the adjacent unit area of the certain unit area. Gather all the adjacent unit areas of the i-th unit area among the N unit areas to form the neighborhood set Q of the i-th unit area i ; Step S202: Obtain the j-th unit area in the neighborhood set Q i , obtain the gray-scale feature value D i of the i-th unit area, the gray-scale feature value D j of the j-th unit area, the Euclidean distance dis ij between the centroids of the i-th unit area and the j-th unit area, and calculate the gray-scale gradient G ij between the i-th unit area and the j-th unit area, G ij = |D i - D j | / dis ij ; Step S203: Traverse the gray-scale gradients between all the unit areas in the neighborhood set Qi and the i-th unit area, and take the unit area corresponding to the maximum gray-scale gradient as the adjacent area of the i-th unit area. The gray-scale gradient between the adjacent area and the i-th unit area is recorded as the edge weight.

[0008] Further, step S300 includes: Step S301: Gather the region feature vectors of all the unit areas in the first target picture to obtain a node feature matrix. According to the order of the unit areas in the node feature matrix, gather all the edge weights of the unit areas to obtain an adjacency matrix; Step S302: Input the node feature matrix and the adjacency matrix into the graph convolutional network, and calculate the feature vectors of each unit region through the graph convolutional network. The feature vectors are denoted as the first feature vectors of the unit regions. By describing the gray-scale features and texture features in the image slices of each unit region, and by comparing the gradient changes of the gray-scale features and texture features between the unit region and its adjacent regions, the connection between the unit regions is established. This solution deeply mines the overall gray-scale distribution and texture features of the image. On the basis of extracting the surface features of the image gray-scale and texture information, such as contrast, energy, homogeneity, and correlation, it further mines the trend of the changes in the gray-scale features and texture features between regions in the picture, providing a basis for identifying the decrease in the clarity of local regions of the image in the subsequent steps.

[0009] Further, step S400 includes: Step S401: Denote the compressed picture of the first target picture as the second target picture. Mark the unit regions of the first target picture in the compressed unit regions in the second target picture, obtain the image code of each unit region in the second target picture. Each unit region's image code corresponds to an image code set. Arrange the image code sets in the order of the unit regions in the second target picture to obtain the first image code sequence. Step S402: Repeat steps S100 to S300 to calculate the feature vectors of each unit region in the second target region. The feature vectors are denoted as the second feature vectors of the unit regions. Step S403: Obtain the first feature vectors and the second feature vectors of each unit region before compression, and denote the distance between the first feature vector and the second feature vector as the difference degree of each unit region. Step S404: Set the difference degree threshold γ, and denote the unit regions with a difference degree greater than the difference degree threshold as abnormal regions. Step S405: Perform image sharpening on all abnormal regions in the second target picture, obtain the image codes of the abnormal regions after the image sharpening process, replace the image codes of the corresponding unit regions in the first image code sequence to obtain the second image code sequence, and feedback the second image code sequence to the relevant management personnel. Detect whether the compressed picture maintains the original gray-scale and texture change features. When the picture is compressed, if the sharpness of some regions decreases and the difference degree of the sharpness changes in different regions of the picture decreases, it will cause the picture to look unclear. Therefore, it is necessary to perform sharpness compensation on the regions with serious sharpness decline to keep the compressed picture have a similar sharpness to the original picture. Meanwhile, the areas of the picture that need to be compensated are screened first, and then the sharpness compensation is performed on the screened areas, reducing the volume of the picture file after compensation. Under the premise of maintaining the gray-scale features and texture features of the picture before and after compression, this solution further controls the volume of the file after compression.

[0010] To better implement the above method, a lightweight video coding system based on image texture and gradient features is also proposed. The system includes: a picture feature management module, a gradient correlation management module, a feature extraction module, and an image sharpening module; The picture feature management module is used to segment the picture and extract the image texture features of each unit area after segmentation. The gradient correlation management module is used to obtain the gradients of the texture features between different unit areas and manage the gradient features between unit areas. The feature extraction module is used to extract the correlation features in the picture through a convolutional network and manage the feature vectors of each unit area in the picture. The image sharpening module is used to evaluate the texture features of the compressed picture, compare them with the feature vectors obtained before compression, and screen out the unit areas with abnormal gradients for sharpening compensation; Furthermore, the picture feature management module includes: a picture segmentation unit, a gray-scale feature management unit, a texture feature management unit, a regional feature vector management unit, and a gray-scale feature value management unit. The picture segmentation unit is used to segment the picture and manage the unit areas in the picture. The gray-scale feature management unit is used to extract and manage the gray-scale features of the picture. The texture feature management unit is used to extract and manage the texture features of the picture. The regional feature vector management unit is used to manage the regional feature vectors of each unit area. The gray-scale feature value management unit is used to manage the gray-scale feature values of each unit area; Furthermore, the gradient correlation management module includes: a domain management unit, a gray-scale gradient management unit, and an adjacent area management unit. The domain management unit is used to perform domain matching on each unit area. The gray-scale gradient management unit is used to calculate the texture gradients between the unit area and each unit area in the domain set. The adjacent area management unit is used to manage the adjacent areas of each unit area and calculate the weights between the unit area and the corresponding adjacent areas; Furthermore, the feature extraction module includes: a feature matrix management unit and a graph convolutional unit; The feature matrix management unit is used to manage the node feature matrix and the adjacent feature matrix. The graph convolutional unit is used to extract the feature vectors of each unit area through a graph convolutional network; Furthermore, the image sharpening module includes: a feature vector management unit, a difference comparison unit, a difference evaluation unit, and a sharpening feedback unit; The eigenvector management unit is used to manage the first eigenvector and the second eigenvector of each unit area, the difference comparison unit is used to manage the difference between the first eigenvector and the second eigenvector, the difference evaluation unit is used to mark the unit area with a difference degree greater than the difference degree threshold as an abnormal area, and the sharpening feedback unit is used to sharpen the abnormal area and feedback the second image coding sequence to relevant management personnel.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Precise feature analysis: The gray scale and texture features of the image slices are described in detail, and the connection is established by comparing the feature gradient changes between regions. The present invention can accurately grasp the overall and local features of the image; 2. Image quality maintenance: Detect the gray scale and texture change features of the compressed picture, and perform sharpening compensation on the areas with serious sharpness decline. It can effectively maintain the sharpness of the compressed picture similar to that of the original picture, ensure the clarity of the image in terms of visual perception, and improve the user experience; 3. File size control: First, screen the areas in the picture that need compensation, and then perform sharpening compensation. While maintaining the image features, the file size of the compensated picture is significantly reduced, achieving a balanced optimization of image quality and file size, which is beneficial for storage and transmission. Description of the Drawings

[0012] Figure 1 It is a schematic structural diagram of the light-load video coding system based on image texture and gradient features of the present invention; Figure 2 It is a schematic flow diagram of the light-load video coding method based on image texture and gradient features of the present invention. Detailed Embodiments

[0013] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0014] Embodiment: As Figure 1 and Figure 2 shown, the present invention provides a technical solution, a light-load video coding method and system based on image texture and gradient features; The method includes: Step S100: Segment the picture before compression processing, and extract the image texture features of each unit area after segmentation; Among them, step S100 includes: Step S101: Obtain the image before compression processing, denote the image as the first target image, divide the first target image into N unit regions, and extract the grayscale feature values and texture feature values of each unit region; Step S102: Compose the grayscale feature values of each unit region into a grayscale feature vector, and compose the texture feature values of each unit region into a texture feature vector; Step S103: Concatenate the grayscale feature vector and the texture feature vector of each unit region into a region feature vector corresponding to each unit region; Step S104: Calculate the grayscale feature values of each unit region, where the grayscale feature values include calculating the contrast of the grayscale image, the uniformity of the grayscale distribution, the similarity of the grayscale distribution, and the correlation of the grayscale values; In the embodiment, first obtain the grayscale image of the first target image, quantize the grayscale in the grayscale image into L quantization intervals to generate the GLCM matrix of the grayscale image, and the row index and column index of the GLCM matrix are both 0 to L-1; Calculate the contrast, energy, homogeneity, and correlation of each unit region through the GLCM matrix; Collect the 4 feature values in the unit region to form a 4-dimensional feature vector, Calculate the LBP feature in the unit region, capture the texture feature of the unit region, output an n-dimensional feature vector, and concatenate it with the above 4-dimensional feature vector to obtain an n+4-dimensional region feature vector. Preferably, when the default 8-neighborhood pixel radius is 1, the LBP feature outputs an n=10-dimensional vector, and the region feature vector at this time is 14-dimensional; In the embodiment, the method for calculating the grayscale feature value is to select one or more of the contrast, energy, homogeneity, and correlation for weighted average to obtain the grayscale feature value corresponding to the unit region.

[0015] Step S200: Obtain the gradient of the texture feature between different unit regions, and use the maximum gradient between the unit region and the adjacent region as the association feature between the unit region and the adjacent region.

[0016] Among them, step S200 includes: Step S201: Obtain the centroid of the unit region. When the distance between the centroid of a certain unit region and the centroid of another unit region is less than the distance threshold, denote the other unit region as the adjacent unit region of the certain unit region, and collect all the adjacent unit regions of the i-th unit region among the N unit regions to form the neighborhood set Q of the i-th unit region i ; Step S202: Obtain the j-th unit region in the neighborhood set Q i and obtain the grayscale feature value D of the i-th unit regioni The gray-scale feature value D of the j-th unit region j The Euclidean distance dis between the centroid of the i-th unit region and the centroid of the j-th unit region ij Calculate the gray-scale gradient G between the i-th unit region and the j-th unit region ij G ij =|D i -D j | / dis ij ; Step S203: Traverse the gray-scale gradients between all unit regions in the neighborhood set Qi and the i-th unit region, and take the unit region corresponding to the maximum gray-scale gradient as the adjacent region of the i-th unit region. The gray-scale gradient between the adjacent region and the i-th unit region is recorded as the edge weight; In the embodiment, to prevent the denominator from approaching 0 during the calculation of the gray-scale gradient, resulting in an overly large calculation result and causing an overflow, the second formula of G can be selected. G ij =|D ij -D i | / (dis j +ε), where ε represents the remainder term. More preferably, ε = e ij +ε), where ε represents the remainder term. More preferably, ε = e -5 , and e is the natural constant.

[0017] Step S300: Extract the associated features in the picture through a convolutional network to obtain the feature vectors of each unit region in the picture; Among them, Step S300 includes: Step S301: Aggregate the regional feature vectors of all unit regions in the first target picture to obtain a node feature matrix. According to the order of the unit regions in the node feature matrix, aggregate the edge weights of all unit regions to obtain an adjacency matrix; Step S302: Input the node feature matrix and the adjacency matrix into the graph convolutional network, and calculate the feature vectors of each unit region through the graph convolutional network. The feature vectors are recorded as the first feature vectors of the unit regions; In the embodiment, the first target picture is segmented into 400 unit regions, and each unit region extracts a 14-dimensional regional feature vector, forming a node feature matrix X with 400 rows and 14 columns. Obtain the row order of the node feature matrix, and according to the order, obtain the edge weights of the adjacent regions of each unit region. The weights of non-adjacent regions are 0. Aggregate all the edge weights and weight values to obtain an adjacency matrix A; Calculate matrix B, B = D -1 / 2 (A + I)D -1 / 2 , where D is the degree matrix, representing the adjacency relationship of the unit regions, and I is the identity matrix; Extract feature vectors through two layers of graph convolutional layers: G 1 = ReLU(B X W 0 ), where ReLU represents the activation function, and W 0 is the weight matrix and is a learnable parameter matrix. The dimension of W 0 is 14×64, and the dimension of the output G 1 is 400×64; G 2 = ReLU(B G 1 W 1 ), W 1 is the weight matrix and is a learnable parameter matrix. The dimension of W 1 is 64×128.

[0018] Step S400: After the picture is compressed, evaluate the texture features of the compressed picture, compare with the feature vectors obtained before compression, and screen out the unit regions with abnormal gradients for sharpening compensation; Among them, step S400 includes: Step S401: Denote the compressed picture of the first target picture as the second target picture. Mark the unit regions of the first target picture in the compressed unit regions in the second target picture, obtain the image encoding of each unit region in the second target picture. Each unit region's image encoding corresponds to an image encoding set. Arrange the image encoding sets in the order of the unit regions in the second target picture to obtain the first image encoding sequence; Step S402: Repeat steps S100 to S300 to calculate the feature vectors of each unit region in the second target region. The feature vectors are denoted as the second feature vectors of the unit regions; Step S403: Obtain the first feature vectors and second feature vectors of each unit region before compression, and denote the distance between the first feature vector and the second feature vector as the difference degree of each unit region; Step S404: Set the difference degree threshold γ, and denote the unit regions with a difference degree greater than the difference degree threshold as abnormal regions; Step S405: Perform image sharpening on all abnormal regions in the second target picture, obtain the image encoding of the abnormal regions after the image sharpening process, replace the image encoding of the corresponding unit regions in the first image encoding sequence to obtain the second image encoding sequence, and feedback the second image encoding sequence to the relevant management personnel; In the embodiment, number the unit regions in the first target picture. After the first target picture is compressed into the second target picture, correspond each unit region in the first target picture to the second target picture, and continue to use the numbers in the first target picture in the second target picture; Obtain the first eigenvector M of the k-th unit region k1 and the second eigenvector M k2 , by calculating the second-order norm of M k1 and M k2 , obtain the distance ΔM between the first eigenvector and the second eigenvector k ; ΔM k =||M k1 -M k2 || 2 Calculate the global difference degree M global , M global =(ΔM 1 +ΔM 2 +ΔM 3 +……+ΔM N ) / N, where N represents the total number of unit regions; Calculate the distances between the first eigenvectors and the second eigenvectors of all unit regions one by one, and calculate the standard deviation σ of the distances. The difference degree threshold γ = M global - 3σ; When outputting the abnormal region, a visualization scheme such as a heat map or a bar chart can be adopted to represent the amount by which the difference degree of the abnormal region exceeds the difference degree threshold.

[0019] The system includes: a picture feature management module, a gradient correlation management module, a feature extraction module, and an image sharpening module; Among them, the picture feature management module is used to segment the picture and extract the image texture features of each unit region after segmentation. The picture feature management module includes: a picture segmentation unit, a grayscale feature management unit, a texture feature management unit, a region feature vector management unit, and a grayscale feature value management unit. Among them, the picture segmentation unit is used to segment the picture and manage the unit regions in the picture. The grayscale feature management unit is used to extract and manage the grayscale features of the picture. The texture feature management unit is used to extract and manage the texture features of the picture. The region feature vector management unit is used to manage the region feature vectors of each unit region. The grayscale feature value management unit is used to manage the grayscale feature values of each unit region; Among them, the gradient correlation management module is used to obtain the gradients of the texture features between different unit regions and manage the gradient features between unit regions. The gradient correlation management module includes: a domain management unit, a grayscale gradient management unit, and an adjacent region management unit. Among them, the domain management unit is used to perform domain matching on each unit region. The grayscale gradient management unit is used to calculate the texture gradients of the unit region and each unit region in the domain set. The adjacent region management unit is used to manage the adjacent regions of each unit region and calculate the weights of the unit region and the corresponding adjacent regions; Among them, the feature extraction module is used to extract the associated features in the picture through a convolutional network and manage the feature vectors of each unit area in the picture. Among them, the feature extraction module includes: a feature matrix management unit and a graph convolution unit. The feature matrix management unit is used to manage the node feature matrix and the adjacency feature matrix, and the graph convolution unit is used to extract the feature vectors of each unit area through a graph convolution network; Among them, the image sharpening module is used to evaluate the texture features of the compressed picture, compare them with the feature vectors obtained before compression, and screen out the unit areas with abnormal gradients for sharpening compensation. Among them, the image sharpening module includes: a feature vector management unit, a difference comparison unit, a difference evaluation unit, and a sharpening feedback unit. Among them, the feature vector management unit is used to manage the first feature vector and the second feature vector of each unit area, the difference comparison unit is used to manage the difference between the first feature vector and the second feature vector, the difference evaluation unit is used to record the unit areas with a difference degree greater than the difference degree threshold as abnormal areas, and the sharpening feedback unit is used to sharpen the abnormal areas and feedback the second image coding sequence to relevant management personnel.

[0020] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A light-load video coding method based on image texture and gradient features, characterized in that: The method comprises the steps of: Step S100: segmenting the image before compression, and extracting image texture features of each unit area after segmentation; Step S200: obtaining the gradient of texture features between different unit regions, and taking the maximum gradient between the unit region and the adjacent region as the correlation feature between the unit region and the adjacent region; Step S300: extracting the associated features in the image through a convolutional network to obtain a feature vector of each unit area in the image; Step S400: After the image is compressed, texture features of the compressed image are evaluated and compared with feature vectors obtained before compression, and unit areas with abnormal gradients are screened out for sharpening compensation.

2. The light-load video encoding method based on image texture and gradient features according to claim 1 is characterized in that: Step S100 includes: Step S101: obtaining an image before compression processing, recording the image as a first target image, dividing the first target image into N unit areas, and extracting grayscale feature values ​​and texture feature values ​​of each unit area; Step S102: The grayscale feature values ​​of each unit area are combined into a grayscale feature vector, and the texture feature values ​​of each unit area are combined into a texture feature vector; Step S103: splicing the grayscale feature vector and the texture feature vector of each unit area into a region feature vector corresponding to each unit area; Step S104: Calculate the grayscale feature value of each unit area, wherein the grayscale feature value includes calculating the contrast of the grayscale image, the uniformity of the grayscale distribution, the similarity of the grayscale distribution, and the correlation of the grayscale value.

3. The light-load video encoding method based on image texture and gradient features according to claim 2 is characterized in that: Step S200 includes: Step S201: Obtain the centroid of a unit area. When the distance between the centroid of a unit area and the centroid of another unit area is less than a distance threshold, the other unit area is recorded as an adjacent unit area of ​​the unit area. All adjacent unit areas of the i-th unit area in the N unit areas are collected to form a domain set Q of the i-th unit area. i ; Step S202: Obtain domain set Q i The j-th unit area in the ith unit area is obtained, and the grayscale feature value D of the i-th unit area is obtained. i , the grayscale eigenvalue D of the jth unit area j , the Euclidean distance dis between the centroid of the i-th unit area and the j-th unit area ij , calculate the gray gradient G between the i-th unit area and the j-th unit area ij , G ij =|D i -D j | / dis ij ; Step S203: Traverse the domain set Q i The grayscale gradients of all unit areas in and the i-th unit area are recorded, and the unit area corresponding to the maximum grayscale gradient is taken as the adjacent area of ​​the i-th unit area, and the grayscale gradient of the adjacent area and the i-th unit area is recorded as the edge weight.

4. The light-load video encoding method based on image texture and gradient features according to claim 3 is characterized in that: Step S300 includes: Step S301: Collecting the regional feature vectors of all unit regions in the first target image to obtain a node feature matrix, and collecting the edge weights of all unit regions according to the order of the unit regions in the node feature matrix to obtain an adjacency matrix; Step S302: Input the node feature matrix and the adjacency matrix into the graph convolutional network, and calculate the feature vector of each unit area through the graph convolutional network. The feature vector is recorded as the first feature vector of the unit area.

5. The light-load video encoding method based on image texture and gradient features according to claim 4 is characterized in that: Step S400 includes: Step S401: Recording the compressed image of the first target image as the second target image, marking the compressed unit area of ​​the unit area of ​​the first target image in the second target image, obtaining the image code of each unit area in the second target image, the image code of each unit area corresponds to an image code set, and arranging the image code set according to the arrangement order of the unit areas in the second target image to obtain a first image code sequence; Step S402: repeating steps S100 to S300 to calculate the feature vector of each unit area in the second target area, where the feature vector is recorded as the second feature vector of the unit area; Step S403: obtaining the first eigenvector and the second eigenvector of each unit area before compression, and recording the distance between the first eigenvector and the second eigenvector as the difference degree of each unit area; Step S404: setting a difference threshold γ, and marking a unit area with a difference greater than the difference threshold as an abnormal area; Step S405: perform image sharpening on all abnormal areas in the second target image, obtain the image code of the abnormal area after image sharpening, replace the image code of the corresponding unit area in the first image code sequence, obtain the second image code sequence, and feed back the second image code sequence to relevant management personnel.

6. A light-load video coding system based on image texture and gradient features, used to execute the light-load video coding method based on image texture and gradient features according to any one of claims 1 to 5, characterized in that: The system includes: Picture feature management module, gradient association management module, feature extraction module and image sharpening module; The image feature management module is used to segment the image and extract the image texture features of each unit area after segmentation. The gradient association management module is used to obtain the gradient of texture features between different unit areas and manage the gradient features between unit areas. The feature extraction module is used to extract the associated features in the image through a convolutional network and manage the feature vectors of each unit area in the image. The image sharpening module is used to evaluate the texture features of the compressed image, compare it with the feature vector obtained before compression, and screen out the unit areas with abnormal gradients for sharpening compensation.

7. The light-load video coding system based on image texture and gradient features according to claim 6, characterized in that: The picture feature management module includes: a picture segmentation unit, a grayscale feature management unit, a texture feature management unit, a regional feature vector management unit and a grayscale feature value management unit; The image segmentation unit is used to segment the image and manage the unit areas in the image. The grayscale feature management unit is used to extract and manage the grayscale features of the image. The texture feature management unit is used to extract and manage the texture features of the image. The regional feature vector management unit is used to manage the regional feature vectors of each unit area. The grayscale feature value management unit is used to manage the grayscale feature values ​​of each unit area.

8. The light-load video coding system based on image texture and gradient features according to claim 6, characterized in that: The gradient association management module includes: a domain management unit, a grayscale gradient management unit and an adjacent area management unit; The domain management unit is used to perform domain matching on each unit area, the grayscale gradient management unit is used to calculate the texture gradient of the unit area and each unit area in the domain set, and the adjacent area management unit is used to manage the adjacent areas of each unit area and calculate the weights of the unit area and the corresponding adjacent areas.

9. The light-load video coding system based on image texture and gradient features according to claim 6, characterized in that: The feature extraction module includes: a feature matrix management unit and a graph convolution unit; The feature matrix management unit is used to manage the node feature matrix and the adjacency feature matrix, and the graph convolution unit is used to extract the feature vector of each unit area through the graph convolution network.

10. The light-load video coding system based on image texture and gradient features according to claim 6, characterized in that: The image sharpening module includes: a feature vector management unit, a difference comparison unit, a difference evaluation unit and a sharpening feedback unit; The feature vector management unit is used to manage the first feature vector and the second feature vector of each unit area, the difference comparison unit is used to manage the difference between the first feature vector and the second feature vector, the difference evaluation unit is used to record the unit area with a difference greater than a difference threshold as an abnormal area, and the sharpening feedback unit is used to sharpen the abnormal area and feed back the second image coding sequence to relevant management personnel.

Images