Method and device for enhancing skin texture of human face in image and video
By acquiring and processing the image texture of the video stream, combining preset texture details, dividing the face area for pixel processing, the problem of unrealistic facial texture enhancement effect in the prior art is solved, and the skin texture enhancement of natural deformation is achieved The effect improves the authenticity and delicateness of the texture of the human face.
Patent Information
- Application Number
- CN202111532483.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-15
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2041-12-15
AI Technical Summary
The prior art enhancement effect of the face skin texture in portrait editing is not realistic and natural enough, and the noise texture cannot naturally produce deformation effect when it deforms or changes in orientation.
By obtaining the image texture of the video stream, determining the key point information of the face, using the pre-trained face preset information to extract the model, dividing the face area into multiple triangle areas, using the fragment shader to perform pixel processing on these areas, combining pre-pre- Send texture details to draw to achieve natural deformation skin texture enhancement.
It achieves a real and delicate skin texture enhancement effect, and can naturally deform the texture when the face is deformed or the direction changes, improving the authenticity and delicateness of the face texture.
Smart Images
Figure CN114187340B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of image processing, and in particular relates to a method and a device for enhancing the texture of human face skin in image videos. Background Art
[0002] With the increasing processing power of mobile terminals and the widespread popularity of image and video editing applications, users' demand for portrait editing is increasing, and they have higher requirements for the delicateness and authenticity of the beautification effect. In addition to conventional editing, filter effects, facial feature adjustment, skin smoothing and other beautification, it is also necessary to beautify and enhance the texture details of the facial skin to improve the texture and overall temperament of the face.
[0003] In the related technology, when users edit portraits, sometimes they need to enhance the skin texture of the face in order to improve the texture and overall effect of the portrait. Currently, the applications with such editing functions on the market can only be achieved by superimposing a random noise texture, although the generation of noise textures is varied and the mixing methods are also different. However, they all have the following defects: on the one hand, the enhanced texture effect is not delicate and realistic enough, and it is impossible to perfectly restore the skin texture of a real face; on the other hand, when the face in the image is stretched or deformed or the face orientation changes, the noise texture cannot produce a natural deformation effect. Summary of the invention
[0004] In view of this, the purpose of the present invention is to overcome the shortcomings of the prior art and provide a method and device for enhancing the skin texture of the face in images and videos, so as to solve the problem that the skin texture details are not realistic and natural enough when the image and video editing in the prior art is processed.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for enhancing the texture of human face skin in an image or video, comprising:
[0006] Acquire the image texture of the video stream, and determine the facial key point information of the face area in the image texture; wherein the facial key point information at least includes: the two-dimensional coordinates of the facial features and the facial contour in the image;
[0007] The face preset information is obtained by extracting the face preset information through the pre-trained face preset information model; wherein the face preset information includes: preset face key point information and preset texture detail map;
[0008] According to the image texture and the facial key point information, the connection relationship between the facial key points in the image texture is calculated, and the facial area in the image texture is divided into a plurality of triangular areas according to the connection relationship; the plurality of triangular areas are pixel-processed using the facial preset information to output a texture-enhanced image.
[0009] Furthermore, the step of acquiring the image texture of the video stream and determining the facial key point information of the facial information in the image texture includes:
[0010] Decoding the video stream to obtain image texture;
[0011] The image texture is input into a pre-trained facial key point detection model to output facial key point information.
[0012] Furthermore, the face preset information is obtained by extracting the face preset information model through pre-training, including:
[0013] Collecting a preset face texture of a face image, marking key points of the preset face texture, and obtaining preset face key point information;
[0014] The preset face texture is extracted according to the preset face key point information to obtain a preset texture detail map.
[0015] Further, the extracting the preset face texture according to the preset face key point information to obtain a preset texture detail map includes:
[0016] Grayscale the preset face texture to obtain a grayscale texture;
[0017] Performing Gaussian filtering on the grayscale texture to obtain a blurred texture;
[0018] Performing filtering processing on the preset face texture to obtain a strong blurred texture;
[0019] Acquire a texture difference between a blurred texture and a strongly blurred texture, and determine the texture difference as a texture detail;
[0020] Performing nonlinear correction on the texture details one by one to obtain a corrected texture;
[0021] The non-face area in the corrected texture is eliminated, and the transparency of the eliminated area is uniformly set to obtain a preset texture detail map.
[0022] Furthermore, before graying the face texture, the method further includes:
[0023] The non-face area in the face image, as well as the eyebrows, eyes, nose, and mouth on the face area are removed, and the transparency of the removed areas is uniformly set.
[0024] Furthermore, the method of using the preset face information to perform pixel processing on multiple triangular areas and outputting a texture-enhanced image includes:
[0025] The preset facial key point information is used as a texture coordinate group, the preset texture detail map is used as an input texture, the facial key point information of the image texture is used as a vertex coordinate group, and the image texture is used as an input texture. The fragment shader is used to draw the multiple triangular areas that are divided one by one.
[0026] Furthermore, the method of using a fragment shader to draw the divided multiple triangular regions one by one includes:
[0027] Determine the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture;
[0028] Pixel calculation is performed using the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture to obtain different color results for drawing.
[0029] Furthermore, the culling area is uniformly set to a gradient transparent color.
[0030] Furthermore, the Delaunay triangulation method is used to segment the face region into a number of non-overlapping triangular regions.
[0031] The embodiment of the present application provides a device for enhancing the texture of human face skin in an image or video, including:
[0032] A face detection module is used to obtain the image texture of the video stream and determine the face key point information of the face area in the image texture; wherein the face key point information at least includes: the two-dimensional coordinates of the facial features and the facial contour in the image;
[0033] A face preset information extraction module is used to obtain face preset information through a pre-trained face preset information extraction model; wherein the face preset information includes: preset face key point information and preset texture detail map;
[0034] A skin texture enhancement module, used to calculate the connection relationship between the key points of the face in the image texture according to the image texture and the key point information of the face, and divide the face area in the image texture into a plurality of triangular areas according to the connection relationship; and perform pixel processing on the plurality of triangular areas using the preset face information;
[0035] Output module, used to output texture-enhanced images.
[0036] The present invention adopts the above technical solution, and the beneficial effects that can be achieved include:
[0037] The present invention provides a method and device for enhancing the skin texture of faces in images and videos. The present invention obtains preset information by pre-detecting key points of real faces and extracting skin texture. When editing faces, the key points of the faces in the input images are identified, and the pre-extracted skin texture is mapped and deformed in combination with the preset information, and then specifically mixed with the original face, thereby obtaining a realistic, delicate image with a naturally deformed skin texture enhancement effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0039] Figure 1 A schematic diagram of the steps of the method for enhancing the texture of human face skin in an image or video according to the present invention;
[0040] Figure 2 It is a flow chart of a method for enhancing the texture of human face skin in an image or video according to the present invention;
[0041] Figure 3 The present invention is a schematic diagram of the structure of a device for enhancing the texture of human face skin in images and videos. DETAILED DESCRIPTION
[0042] To make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.
[0043] A specific method and device for enhancing the skin texture of a face in an image or video provided in an embodiment of the present application is described below in conjunction with the accompanying drawings.
[0044] like Figure 1 As shown, the method for enhancing the skin texture of a face in an image or video provided in an embodiment of the present application includes:
[0045] S101, acquiring an image texture of a video stream, and determining facial key point information of a face area in the image texture; wherein the facial key point information at least includes: two-dimensional coordinates of facial features and facial contours in the image;
[0046] This application first obtains the video to be processed, obtains the image texture of the video stream, determines the face area in the image texture, and determines the face key point information of the face area. The face key point data should at least include the two-dimensional coordinates of the facial features and facial contours in the image. Others may also be included, which are not limited in this application. It should be noted that the higher the fit between the face key point data determined in this application and the real input face, the greater the number, and the more refined the output effect after the subsequent processing of the face skin texture. The specific model and the number of key points used can also be flexibly set according to the actual situation. The model can be implemented using existing technology, and this application will not go into details here.
[0047] S102, obtaining preset face information through a pre-trained face preset information extraction model; wherein the preset face information includes: preset face key point information and preset texture detail map;
[0048] In this application, a preset face information extraction model is trained in advance, and the model can mark key points of the face texture in the input image and extract the texture of the face texture.
[0049] S103, calculating the connection relationship between the facial key points in the image texture according to the image texture and the facial key point information, dividing the facial area in the image texture into a plurality of triangular areas according to the connection relationship; performing pixel processing on the plurality of triangular areas using the facial preset information, and outputting a texture-enhanced image.
[0050] The working principle of the method for enhancing the texture of human face skin texture in image and video provided by the present application is as follows: firstly, the video to be processed is obtained, the image texture of the video stream is obtained, the human face area in the image texture is determined, and the human face key point information of the human face area is determined, and the human face key point data should at least include the two-dimensional coordinates of the facial features and the facial contour in the image. According to the image texture and the human face key point information, the connection relationship between the human face key points in the image texture is calculated, and the human face area in the image texture is divided into multiple triangular areas according to the connection relationship; the human face preset information is used to perform pixel processing on the multiple triangular areas, and a texture-enhanced image is output.
[0051] In some embodiments, the acquiring of the image texture of the video stream and the determining of the facial key point information of the facial information in the image texture include:
[0052] Decoding the video stream to obtain image texture;
[0053] The image texture is input into a pre-trained facial key point detection model to output facial key point information.
[0054] Specifically, the present application obtains the image texture of the video stream and outputs the facial information in the image. Specifically, the decoded video frame image or image texture input by the user is detected using a facial key point detection model based on deep learning, and the facial key point information is output when and only when the facial key point detection model detects the facial key point information. Preferably, the facial key point information should at least include the two-dimensional coordinates of the facial features and facial contours in the image. It can be understood that the higher the fit between the key points and the real input face and the greater the number, the more refined the output effect after subsequent processing. The specific detection model and the number of key points used can also be flexibly set according to the actual situation, and this application does not limit it here.
[0055] In some embodiments, obtaining the preset face information by extracting the pre-trained preset face information model includes:
[0056] Collecting a preset face texture of a face image, marking key points of the preset face texture, and obtaining preset face key point information;
[0057] The preset face texture is extracted according to the preset face key point information to obtain a preset texture detail map.
[0058] Preferably, the step of extracting the preset face texture according to the preset face key point information to obtain a preset texture detail map includes:
[0059] Grayscale the preset face texture to obtain a grayscale texture;
[0060] Performing Gaussian filtering on the grayscale texture to obtain a blurred texture;
[0061] Performing filtering processing on the preset face texture to obtain a strong blurred texture;
[0062] Acquire a texture difference between a blurred texture and a strongly blurred texture, and determine the texture difference as a texture detail;
[0063] Performing nonlinear correction on the texture details one by one to obtain a corrected texture;
[0064] The non-face area in the corrected texture is eliminated, and the transparency of the eliminated area is uniformly set to obtain a preset texture detail map.
[0065] Preferably, before graying the facial texture, the method further includes:
[0066] The non-face area in the face image, as well as the eyebrows, eyes, nose, and mouth on the face area are removed, and the transparency of the removed areas is uniformly set.
[0067] Specifically, as a specific implementation, the training process of the face preset information extraction model includes the following steps:
[0068] 1. Collect a preset face texture, that is, a frame image. The collected preset face texture is preferably a frontal face, a large picture, regular facial features, strong skin texture, and no or few skin defects. The source of the face texture can be collected from a real face, or it can be obtained by modeling and mapping using modeling software and other technical means. Preferably, if the collected face data has obvious defects or the skin texture is not strong, automatic or manual means can be used to repair and enhance the defects.
[0069] 2. Label the key points of the input face texture. Preferably, a face key point detection model based on machine learning can be used for detection. The pixel width and height of the face area and a set of ordered normalized two-dimensional coordinates of the face key points are obtained, such as [width, height, x1, y1, x2, y2, x3, y3, ...], which are pre-stored as the first part of the face preset information.
[0070] 3. Extract the input face texture. The process of this step is as follows:
[0071] 1) Combined with the preset facial key point information in the above step 2, remove the area outside the face, eyebrows, eyes, nose, mouth, etc., and make the color after removal transparent. This serves as the input source for the next step.
[0072] 2) Grayscale the input preset face texture pixel by pixel to obtain grayscale texture S1. Preferably, for an image without gamma correction, the grayscale calculation formula used in the present invention is: gray = red * 0.299 + green * 0.587 + blue * 0.114. For an image after gamma correction, the grayscale calculation formula is: gray = red * 0.2126 + green * 0.7152 + blue * 0.0755.
[0073] 3) Perform a Gaussian filter on S1 with a convolution kernel size of 3x3 and sigma = 1.0 to obtain a slightly blurred texture B1. Practice has shown that small-scale Gaussian filtering can smooth the slight noise and outliers that originally exist in the image while retaining skin texture information, thereby improving the quality of the final extracted skin texture.
[0074] 4) Taking the input face texture as S2, pre-set large-scale filtering of S2 to obtain a strongly blurred texture B2. Preferably, considering the ratio of the width of the face skin detail texture to the face width in reality, the filtering degree should not be too large or too small. In the present invention, the face area pixel width*0.0834 obtained in the above step 2 is used as the filtering radius R, and the convolution kernel size of the final Gaussian filter is 2*R+1, sigma=R*0.334.
[0075] 5) Subtract texture B2 from texture B1 to obtain texture detail Diff = B1 – B2, where Diff represents the texture detail.
[0076] 6) Perform nonlinear correction on Diff pixel by pixel to obtain the corrected texture feature. The correction formula is:
[0077]
[0078] This step can effectively suppress the parts with too much detail in the Diff texture, avoiding the problem of too much texture detail when mixing in subsequent processes. At the same time, the difference is normalized to the range of 0 to 1, which is convenient for storage and use in actual production.
[0079] 7) Eliminate the areas outside the face in the corrected texture Feature, as well as the eyebrows, eyes, nose, mouth and other parts of the face, and make the colors uniformly transparent after elimination. Preferably, in this step, a gradient transparency can be set for the transparency of the eliminated edge to avoid the problem of unnatural texture transition in the result output by the face skin texture enhancement module.
[0080] 8) The corrected texture Feature is stored as the second part of the preset information, that is, the preset texture detail map.
[0081] Preferably, in the present application, the rejection areas are uniformly set to a gradient transparent color.
[0082] In some embodiments, the using the preset face information to perform pixel processing on multiple triangular areas and outputting a texture-enhanced image includes:
[0083] The preset facial key point information is used as a texture coordinate group, the preset texture detail map is used as an input texture, the facial key point information of the image texture is used as a vertex coordinate group, and the image texture is used as an input texture. The fragment shader is used to draw the multiple triangular areas that are divided one by one.
[0084] Preferably, the method of using a fragment shader to draw the divided multiple triangular regions one by one includes:
[0085] Determine the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture;
[0086] Pixel calculation is performed using the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture to obtain different color results for drawing.
[0087] As a specific implementation, the face preset extracted in the face preset information extraction module is loaded, and the face preset should at least include: preset face key point information and preset texture detail map;
[0088] The original video or image texture frame to be processed and the detected face key point information of the image texture frame are loaded. The connection relationship between the key points in the original texture frame is calculated using the triangulation method, and the face area is divided into a plurality of non-overlapping triangular areas. Preferably, the present invention will use the Delaunay triangulation method.
[0089] The key point information of the preset face is used as the texture coordinate group, the texture detail map of the preset face information is used as the input texture B1, the key point information of the face of the texture frame is used as the vertex coordinate group, and the image texture frame is used as the input texture B2. The divided triangle areas are drawn one by one using the GPU. The fragment shader calculation formula is as follows:
[0090] s2: current pixel color of the original image, 1x3 dimensional vector;
[0091] s1: preset face texture current pixel color brightness, 1x3 dimensional vector;
[0092] a1: The current transparency of the preset face texture, 1x3 dimensional vector;
[0093] resultRed=s2.red+s1.red*2.0-1.0
[0094] resultGreen=s2.green+s1.green*2.0-1.0
[0095] resultBlue=s2.blue+s1.blue*2.0-1.0
[0096] resultColor=s2*(1.0-a1)+(resultRed, resultGreen, resultBlue)*a1
[0097] Finally, all pixels of the video image frame are processed, and the final result of skin texture enhancement is obtained.
[0098] Through the above technical solution, this application can more perfectly and delicately restore the real skin texture and texture details by extracting texture based on real or modeled faces, compared with the existing technology. The present invention deforms the preset extracted texture based on the face detected by key points and then mixes it with the input image. Thus, the corresponding deformation and stretching of the skin texture is generated according to different face angles and orientations, which can effectively simulate the skin effect in a real environment.
[0099] like Figure 3 As shown, the embodiment of the present application provides a device for enhancing the texture of human face skin in an image or video, comprising:
[0100] The face detection module 301 is used to obtain the image texture of the video stream and determine the face key point information of the face area in the image texture; wherein the face key point information at least includes: the two-dimensional coordinates of the facial features and the facial contour in the image;
[0101] The face preset information extraction module 302 is used to obtain the face preset information through the pre-trained face preset information extraction model; wherein the face preset information includes: preset face key point information and preset texture detail map;
[0102] The skin texture enhancement module 303 is used to calculate the connection relationship between the key points of the face in the image texture according to the image texture and the key point information of the face, and divide the face area in the image texture into multiple triangular areas according to the connection relationship; and perform pixel processing on the multiple triangular areas using the preset face information;
[0103] The output module 304 is used to output the texture-enhanced image.
[0104] The working principle of the device for enhancing the skin texture of the face in an image video provided by the present application is as follows: a face detection module 301 obtains the image texture of a video stream and determines the face key point information of the face area in the image texture; wherein the face key point information at least includes: two-dimensional coordinates of the facial features and the facial contour in the image; a face preset information extraction module 302 obtains face preset information through a pre-trained face preset information extraction model; wherein the face preset information includes: preset face key point information and preset texture detail map; a skin texture enhancement module 303 calculates the connection relationship between the face key points in the image texture according to the image texture and the face key point information, and divides the face area in the image texture into multiple triangular areas according to the connection relationship; pixel processing is performed on multiple triangular areas using the face preset information; and an output module 304 outputs a texture-enhanced image.
[0105] In summary, the present invention provides a method and device for enhancing the texture of skin texture of human face in image video, the method comprising obtaining the image texture of a video stream, determining the face key point information of the face area in the image texture; wherein the face key point information at least comprises: the two-dimensional coordinates of the facial features and the facial contour in the image; obtaining the face preset information through a pre-trained face preset information extraction model; wherein the face preset information comprises: preset face key point information and preset texture detail map; according to the image texture and the face key point information, calculating the connection relationship between the face key points in the image texture, dividing the face area in the image texture into a plurality of triangular areas according to the connection relationship; using the face preset information, performing pixel processing on the plurality of triangular areas, and outputting a texture enhanced image. The present invention obtains preset information by pre-detecting key points and extracting skin texture of a real face, and when performing face editing, identifies the face key points of the input image, maps and deforms the pre-extracted skin texture in combination with the preset information, and then performs a specific mixing with the original face, thereby obtaining a real, delicate image with a natural deformation skin texture enhancement effect.
[0106] It can be understood that the method embodiment provided above corresponds to the device embodiment described above, and the corresponding specific contents can be referenced to each other and will not be repeated here.
[0107] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.
[0108] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0109] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction method, which is implemented in the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0110] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0111] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. A method for enhancing the skin texture of a face in an image or video, characterized in that: include: Acquire the image texture of the video stream, and determine the facial key point information of the face area in the image texture; wherein the facial key point information at least includes: the two-dimensional coordinates of the facial features and the facial contour in the image; The face preset information is obtained by extracting the face preset information through the pre-trained face preset information model; wherein the face preset information includes: preset face key point information and preset texture detail map; According to the image texture and the facial key point information, the connection relationship between the facial key points in the image texture is calculated, and the facial region in the image texture is divided into a plurality of triangular regions according to the connection relationship; the plurality of triangular regions are pixel-processed using the facial preset information to output a texture-enhanced image; The method of using the preset face information to perform pixel processing on multiple triangular areas and outputting a texture-enhanced image includes: The preset facial key point information is used as a texture coordinate group, the preset texture detail map is used as an input texture, the facial key point information of the image texture is used as a vertex coordinate group, and the image texture is used as an input texture, and the fragment shader is used to draw the multiple triangular areas that are divided one by one; The method of using a fragment shader to draw the divided multiple triangular areas one by one includes: Determine the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture; Pixel calculation is performed using the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture to obtain different color results for drawing.
2. The method according to claim 1, characterized in that The step of acquiring the image texture of the video stream and determining the facial key point information of the facial information in the image texture includes: Decoding the video stream to obtain image texture; The image texture is input into a pre-trained facial key point detection model to output facial key point information.
3. The method according to claim 1, characterized in that The method of extracting the face preset information through the pre-trained face preset information model to obtain the face preset information includes: Collecting a preset face texture of a face image, marking key points of the preset face texture, and obtaining preset face key point information; The preset face texture is extracted according to the preset face key point information to obtain a preset texture detail map.
4. The method according to claim 3, characterized in that The step of extracting the preset face texture according to the preset face key point information to obtain a preset texture detail map includes: Grayscale the preset face texture to obtain a grayscale texture; Performing Gaussian filtering on the grayscale texture to obtain a blurred texture; Performing filtering processing on the preset face texture to obtain a strong blurred texture; Acquire a texture difference between a blurred texture and a strongly blurred texture, and determine the texture difference as a texture detail; Performing nonlinear correction on the texture details one by one to obtain a corrected texture; The non-face area in the corrected texture is eliminated, and the transparency of the eliminated area is uniformly set to obtain a preset texture detail map.
5. The method according to claim 4, characterized in that Before graying the face texture, it also includes: The non-face area in the face image, as well as the eyebrows, eyes, nose, and mouth on the face area are removed, and the transparency of the removed areas is uniformly set.
6. The method according to claim 4 or 5, characterized in that: Set the culling area to a uniform gradient transparent color.
7. The method according to claim 1, characterized in that The Delaunay triangulation method is used to divide the face area into several non-overlapping triangular areas.
8. A device for enhancing the texture of human face skin in an image or video, characterized in that: include: A face detection module is used to obtain the image texture of the video stream and determine the face key point information of the face area in the image texture; wherein the face key point information at least includes: the two-dimensional coordinates of the facial features and the facial contour in the image; A face preset information extraction module is used to obtain face preset information through a pre-trained face preset information extraction model; wherein the face preset information includes: preset face key point information and preset texture detail map; A skin texture enhancement module, used to calculate the connection relationship between the key points of the face in the image texture according to the image texture and the key point information of the face, and divide the face area in the image texture into a plurality of triangular areas according to the connection relationship; and perform pixel processing on the plurality of triangular areas using the preset face information; An output module, used for outputting a texture-enhanced image; The method of using the preset face information to perform pixel processing on multiple triangular areas and outputting a texture-enhanced image includes: The preset facial key point information is used as a texture coordinate group, the preset texture detail map is used as an input texture, the facial key point information of the image texture is used as a vertex coordinate group, and the image texture is used as an input texture, and the fragment shader is used to draw the multiple triangular areas that are divided one by one; The method of using a fragment shader to draw the divided multiple triangular areas one by one includes: Determine the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture; Pixel calculation is performed using the current pixel color of the image texture, the current pixel color brightness of the preset face texture, and the current transparency of the preset face texture to obtain different color results for drawing.