Animation generation system based on human body structure

Abstract

The invention relates to the technical field of animation generation, and particularly discloses a human body structure-based animation generation system, which comprises a data acquisition module, a data processing module, an action library module, an action matching module and an animation generation module, the method comprises the following steps: acquiring three-dimensional coordinates of joint points corresponding to different individuals; the human body joint point model is processed to generate a standardized human body joint point model; selecting corresponding virtual walking show actions and virtual walking show postures according to the input garment style; the corresponding 3D clothing show animation is generated based on the virtual show actions and the virtual show postures, so that the accuracy and universality of the human body model are improved, and the flexibility and expressive force of clothing display are enhanced.

Description

Technical Field

[0001] The invention belongs to the technical field of animation generation, and relates to an animation generation system based on human body structure. Background Art

[0002] With the rapid development of science and technology and the changing consumption patterns, the clothing display mode of traditional fashion shows is facing unprecedented challenges. As a new way of clothing display in the "post-Internet era", 3D virtual fashion shows have gradually become an important means of clothing display and marketing with the rise of the metaverse. The realization of existing virtual fashion shows is mainly based on special effects production and animation simulation. The cost of special effects production is high, and animation simulation cannot meet the needs of the diversity of models' virtual catwalk movements.

[0003] The current implementation of 3D virtual fashion shows is mainly based on special effects production and animation simulation. The cost of special effects production is high, while animation simulation cannot meet the needs of model virtual catwalk movement diversity. In order to realize the generation of model virtual catwalk movements, it is first necessary to perform 3D posture estimation on the model movements in the catwalk video. This process is achieved by detecting the position information of more than a dozen joints of the human body. However, each person's human joints are highly flexible, and these joints are easily affected by factors such as height and weight, resulting in individual differences, which leads to complex structures and relationships between human joints, and increases the difficulty of joint prediction and connection; At the same time, when different styles of clothing are displayed, models will show different catwalk styles. However, the current method is to match different styles of clothing through virtual catwalk movements of models of the same style, which cannot improve the clothing display effect. Summary of the invention

[0004] In view of the above problems existing in the prior art, the present invention provides an animation generation system based on human body structure to solve the above technical problems.

[0005] In order to achieve the above purpose and other purposes, the technical solution adopted by the present invention is as follows: On one hand, the present invention provides an animation generation system based on human body structure, including a data acquisition module, a data processing module, an action library module, an action matching module and an animation generation module, wherein the above modules are connected by wired and / or wireless connection to realize data transmission between the modules; Data acquisition module: used to collect the three-dimensional coordinates of each joint point corresponding to each difference individual; Data processing module: connected to the data acquisition module, used to process the three-dimensional coordinates of each joint point corresponding to each different individual, and generate a standardized human joint point model; Action library module: used to store various virtual catwalk actions; An action matching module, connected to the data processing module and the action library module, for selecting corresponding virtual catwalk actions and virtual catwalk postures according to the input clothing style; An animation generation module is connected to the action matching module and generates corresponding 3D clothing catwalk animation based on virtual catwalk actions and virtual catwalk postures; Furthermore, the overall operation logic of the data processing module is: Select each key node of the human body from each joint, thereby obtaining the three-dimensional coordinates of each key node of the human body corresponding to each difference individual, and record it as , i is the number of each difference individual, i=1,2,...A, j is the number of each key point of the human body, j=1,2,...J; Select the left shoulder node and the right shoulder node from each key node of the human body, obtain the three-dimensional coordinates of the left shoulder node and the right shoulder node corresponding to each difference individual, and then calculate the direction vector of the left shoulder node and the right shoulder node corresponding to each difference individual , is the three-dimensional coordinate of the left shoulder node corresponding to the i-th difference individual, is the three-dimensional coordinate of the right shoulder node corresponding to the i-th difference individual; The direction vectors of the left shoulder node and the right shoulder node corresponding to each difference individual Rotate to target direction , where the specific calculation steps of the rotation matrix R are as follows: Calculate the axis of rotation , is the norm symbol; calculate the rotation angle ; The rotation matrix is ​​calculated from this , The rotation axis The cross product matrix of I is the identity matrix, are the rotation axes corresponding to the x-axis, y-axis and z-axis respectively; Rotate the key nodes of each human body corresponding to each difference individual to the target direction to obtain the rotated three-dimensional coordinates of the key nodes of each human body corresponding to each difference individual .

[0006] Furthermore, the overall operation logic of the data processing module also includes the following steps: Obtain the rotated three-dimensional coordinates of each key node of the human body corresponding to each difference individual, calculate the mean value thereof, and obtain the mean value of the rotated three-dimensional coordinates corresponding to each key node of the human body; The standard deviation of the three-dimensional coordinates of the rotation of each key node of the human body corresponding to each difference individual is calculated synchronously to obtain the standard deviation of the three-dimensional coordinates of the rotation of each key node of the human body; The rotated three-dimensional coordinates of each key node of the human body corresponding to each difference individual are standardized using the mean and standard deviation of the rotated three-dimensional coordinates of each key node of the human body to obtain the standardized three-dimensional coordinates of each key node of the human body corresponding to each difference individual. , is the mean of the rotated three-dimensional coordinates corresponding to the jth human key node, is the standard deviation of the rotated three-dimensional coordinates corresponding to the jth human key node; The standardized three-dimensional coordinates of each key human body node corresponding to each different individual are imported into a predefined human body joint point model to generate the human body joint point model of each different individual.

[0007] Furthermore, the standardized human joint point model is generated by: Obtain the three-dimensional coordinates of each joint in the human joint point model of each difference individual, calculate the distance between each pair of adjacent nodes in each difference individual based on the three-dimensional coordinates, and then calculate the mean value of the distance between each pair of adjacent nodes; The mean distance between each pair of adjacent nodes is imported into the reference human joint point model to generate a standardized human joint point model.

[0008] Furthermore, the corresponding virtual catwalk action is selected according to the input clothing style. The specific operation logic is as follows: Extract features from the input clothing style and obtain the values ​​of each feature corresponding to the input clothing style ; And extract features from the various virtual catwalk actions stored in the action library module to obtain the values ​​of each feature corresponding to each virtual catwalk action ; The feature similarity between the input clothing style corresponding features and the features corresponding to each virtual catwalk action is calculated. , c is the number of each feature, c=1,2,...B, B is the total number of features, are the values ​​of the first and second features corresponding to the input clothing style, are the values ​​of the first feature and the second feature corresponding to the f-th virtual catwalk action, respectively, and f is the number of each virtual catwalk action; The feature similarities between the input clothing style corresponding features and the features corresponding to each virtual catwalk action are arranged in descending order, and the virtual catwalk action with the highest feature similarity is selected as the virtual catwalk action of the input clothing style.

[0009] Furthermore, the values ​​of each feature corresponding to the input clothing style are obtained. The specific acquisition logic is: Collect M clothing images of input clothing styles, adjust all clothing images to a uniform size and normalize the pixel values ​​of all clothing images; Calculate the color histogram of all clothing images, calculate the histogram of the three RGB channels respectively, and convert the image from RGB space to HSV color space to extract color features; Use LBP algorithm to extract texture features of all clothing images; Use the Canny edge detection algorithm to extract the contour features of all clothing images; Input the preprocessed image into the pretrained model and extract the output of the last fully connected layer as the feature vector; All extracted features are combined into a comprehensive feature vector as the value of each feature corresponding to the input clothing style.

[0010] Furthermore, the virtual catwalk posture is divided into virtual catwalk step length and virtual catwalk speed.

[0011] Furthermore, the corresponding virtual catwalk posture is selected, and the specific selection logic is: According to the input clothing style, calculate the input clothing style coefficient ; According to the standardized human joint point model, the leg length L corresponding to the standardized human joint point model is obtained; The virtual catwalk step length corresponding to the input clothing style is calculated ; Calculate the virtual catwalk speed corresponding to the input clothing style , The benchmark catwalk speed is set.

[0012] Further, the input clothing style coefficient is calculated , the specific calculation logic is: Obtain the clothing type score, brand style score, and display intention score corresponding to the input clothing style, and set the calculation weights of the clothing type score, brand style score, and display intention score respectively; Calculate the clothing style coefficient for the input , are the calculation weights of clothing type score, brand style score and display intention score, respectively. They are the clothing type score, brand style score, and display intention score corresponding to the input clothing style.

[0013] Another aspect of the present invention provides a method for generating an animation based on human body structure, comprising the following steps: Collect the three-dimensional coordinates of each joint point corresponding to each difference individual; Process the three-dimensional coordinates of each joint point of each individual to generate a standardized human joint point model; Select corresponding virtual catwalk movements and virtual catwalk postures according to the input clothing style; Generate corresponding 3D clothing catwalk animation based on virtual catwalk movements and virtual catwalk postures.

[0014] As described above, the human body structure-based animation generation system provided by the present invention has at least the following beneficial effects: The animation generation system based on human body structure provided by the present invention includes a data acquisition module, a data processing module, an action library module, an action matching module and an animation generation module; by collecting the three-dimensional coordinates of each joint point corresponding to each different individual and processing them, a standardized human joint point model is generated, which helps to create a universal and highly accurate human skeleton foundation. This can not only adapt to the changes in different body shapes and postures, but also ensure accurate simulation of the natural movement of the human body in a virtual environment.

[0015] The embodiment of the present invention selects corresponding virtual catwalk movements and virtual catwalk postures according to the input clothing styles, so that the virtual models can show the best visual effects when showing different clothing. In this way, the design characteristics and functions of clothing can be better reflected.

[0016] (3) The embodiments of the present invention improve the accuracy and versatility of the human body model, enhance the flexibility and expressiveness of clothing display, and significantly reduce the cost and time investment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0018] Figure 1 It is a schematic diagram of the connection of various modules of the system of the present invention.

[0019] Figure 2 It is a schematic diagram of the connection of each step of the method of the present invention. DETAILED DESCRIPTION

[0020] The above contents in combination with the implementation of the present invention are merely examples and explanations of the concept of the present invention. The technical personnel in the relevant technical field may make various modifications or supplements to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they shall all fall within the protection scope of the present invention. Example

[0021] See also Figure 1 As shown, the animation generation system based on the human body structure includes a data acquisition module, a data processing module, an action library module, an action matching module and an animation generation module, and the above modules are connected by wired and / or wireless connection to realize data transmission between the modules; Data acquisition module: used to collect the three-dimensional coordinates of each joint point corresponding to each difference individual; Data processing module: connected to the data acquisition module, used to process the three-dimensional coordinates of each joint point corresponding to each different individual, and generate a standardized human joint point model; In a preferred technical solution of the present application, the overall operation logic of the data processing module is: Select each key human body node from each joint, where the key human body nodes include but are not limited to the head, shoulders, elbows, wrists, hips, knees and ankles, thereby obtaining the three-dimensional coordinates of each key human body node corresponding to each different individual, which are recorded as , i is the number of each difference individual, i=1,2,...A, j is the number of each key point of the human body, j=1,2,...J; Select the left shoulder node and the right shoulder node from each key node of the human body, obtain the three-dimensional coordinates of the left shoulder node and the right shoulder node corresponding to each difference individual, and then calculate the direction vector of the left shoulder node and the right shoulder node corresponding to each difference individual , is the three-dimensional coordinate of the left shoulder node corresponding to the i-th difference individual, is the three-dimensional coordinate of the right shoulder node corresponding to the i-th difference individual; The direction vectors of the left shoulder node and the right shoulder node corresponding to each difference individual Rotate to target direction , where the specific calculation steps of the rotation matrix R are as follows: Calculate the axis of rotation , is the norm symbol; calculate the rotation angle ; The rotation matrix is ​​calculated from this , The rotation axis The cross product matrix of I is the identity matrix, which is a matrix whose diagonal elements are all 1 and other elements are all 0. are the rotation axes corresponding to the x-axis, y-axis and z-axis respectively; Rotate the key nodes of each human body corresponding to each difference individual to the target direction to obtain the rotated three-dimensional coordinates of the key nodes of each human body corresponding to each difference individual .

[0022] In a preferred technical solution of the present application, the overall operation logic of the data processing module further includes the following steps: Obtain the rotated three-dimensional coordinates of each key node of the human body corresponding to each difference individual, calculate the mean value thereof, and obtain the mean value of the rotated three-dimensional coordinates corresponding to each key node of the human body; The standard deviation of the three-dimensional coordinates of the rotation of each key node of the human body corresponding to each difference individual is calculated synchronously to obtain the standard deviation of the three-dimensional coordinates of the rotation of each key node of the human body; The rotated three-dimensional coordinates of each key node of the human body corresponding to each difference individual are standardized using the mean and standard deviation of the rotated three-dimensional coordinates of each key node of the human body to obtain the standardized three-dimensional coordinates of each key node of the human body corresponding to each difference individual. , is the mean of the rotated three-dimensional coordinates corresponding to the jth human key node, is the standard deviation of the three-dimensional coordinates of the jth human key node; The standardized three-dimensional coordinates of each key human body node corresponding to each different individual are imported into a predefined human body joint point model to generate a human body joint point model of each different individual.

[0023] In a preferred technical solution of the present application, the standardized human joint point model is generated, and the specific generation process is as follows: Obtain the three-dimensional coordinates of each joint in the human joint point model of each difference individual, calculate the distance between each pair of adjacent nodes in each difference individual based on the three-dimensional coordinates, and then calculate the mean value of the distance between each pair of adjacent nodes; The mean distance between each pair of adjacent nodes is imported into the reference human joint point model to generate a standardized human joint point model.

[0024] Action library module: used to store various virtual catwalk actions; An action matching module, connected to the data processing module and the action library module, for selecting corresponding virtual catwalk actions and virtual catwalk postures according to the input clothing style; In the preferred technical solution of the present application, the corresponding virtual catwalk action is selected according to the input clothing style, and the specific operation logic is: Extract features from the input clothing style and obtain the values ​​of each feature corresponding to the input clothing style ; And extract features from the various virtual catwalk actions stored in the action library module to obtain the values ​​of each feature corresponding to each virtual catwalk action ; The feature similarity between the input clothing style corresponding features and the features corresponding to each virtual catwalk action is calculated. , c is the number of each feature, c=1,2,...B, B is the total number of features, are the values ​​of the first and second features corresponding to the input clothing style, are the values ​​of the first feature and the second feature corresponding to the f-th virtual catwalk action, respectively, and f is the number of each virtual catwalk action; The feature similarities between the input clothing style corresponding features and the features corresponding to each virtual catwalk action are arranged in descending order, and the virtual catwalk action with the highest feature similarity is selected as the virtual catwalk action of the input clothing style.

[0025] In a preferred technical solution of the present application, the values ​​of the features corresponding to the input clothing style are obtained, and the specific acquisition logic is: Collect M clothing images of input clothing styles, adjust all clothing images to a uniform size and normalize the pixel values ​​of all clothing images; Calculate the color histogram of all clothing images, calculate the histogram of the three RGB channels respectively, and convert the image from RGB space to HSV color space to extract color features; Use LBP algorithm to extract texture features of all clothing images; Use the Canny edge detection algorithm to extract the contour features of all clothing images; Input the preprocessed image into the pretrained model and extract the output of the last fully connected layer as the feature vector; All extracted features are combined into a comprehensive feature vector as the value of each feature corresponding to the input clothing style.

[0026] In a preferred technical solution of the present application, each virtual catwalk action is split into frames, and noise in the video is removed using a filter or other method; Extract the changes of key points in time to form time series data; analyze the relative positions and angles between key points to extract spatial features; calculate the two dynamic features of speed and acceleration; analyze the changes of joint angles to extract gait features; evaluate the symmetry of left and right body movements; analyze the periodic characteristics of movements, such as step frequency and stride length; Finally, a machine learning algorithm is used to identify the characteristics of each virtual catwalk action.

[0027] In the preferred technical solution of the present application, the virtual catwalk posture is divided into catwalk step length and catwalk speed.

[0028] In the preferred technical solution of this application, the corresponding virtual catwalk posture is selected, and the specific selection logic is: According to the input clothing style, calculate the input clothing style coefficient ; According to the standardized human joint point model, the leg length L corresponding to the standardized human joint point model is obtained; The virtual catwalk step length corresponding to the input clothing style is calculated ; Calculate the virtual catwalk speed corresponding to the input clothing style , The benchmark catwalk speed is set.

[0029] In the preferred technical solution of this application, the input clothing style coefficient is calculated , the specific calculation logic is: Obtain the clothing type score, brand style score, and display intention score corresponding to the input clothing style, and set the calculation weights of the clothing type score, brand style score, and display intention score respectively; Clothing type: The value is between 0 and 1, where 0 represents extremely elegant clothing (such as a haute couture dress) and 1 represents extremely dynamic clothing (such as sportswear); clothing type can be calculated by weighting fabric type, tailoring style, color pattern, and purpose: Fabric type: The value is between 0 and 1, 0 represents extremely elegant fabrics (such as silk), and 1 represents extremely dynamic fabrics (such as sports fabrics); Tailoring style: The value is between 0 and 1, 0 represents elegant and complex tailoring (such as dresses), and 1 represents simple and dynamic tailoring (such as sportswear); Color and pattern: The value can be between 0 and 1, 0 represents elegant colors and patterns (such as monochrome and plain), and 1 represents dynamic colors and patterns (such as bright colors and complex patterns); Usage: The value is between 0 and 1, 0 represents clothing for formal occasions, and 1 represents clothing for sports or casual occasions.

[0030] Brand style: The value is between 0 and 1, where 0 means the brand is more elegant and 1 means the brand is more dynamic. Brand style can be calculated by weighting brand history, design style, target market and advertising style: Brand history: The value is between 0 and 1, where 0 indicates a long history and classic design, and 1 indicates an emerging brand and avant-garde design. Design style: The value is between 0 and 1, where 0 indicates classic design and 1 indicates avant-garde design. Target market: The value is between 0 and 1, where 0 indicates the high-end market and 1 indicates the young or mass market. Advertising style: The value is between 0 and 1, where 0 indicates an elegant advertising style and 1 indicates a dynamic or avant-garde advertising style.

[0031] Display intention: The value is between 0 and 1, where 0 means the display intention is biased towards detail display, and 1 means the display intention is biased towards dynamic display. Display intention can be calculated by weighting the display purpose, target audience, display content, and display form: Display purpose: The value is between 0 and 1, where 0 represents educational purpose and 1 represents entertainment purpose.

[0032] Target audience: The value is between 0 and 1, where 0 represents children and 1 represents adults.

[0033] Display content: The value is between 0 and 1, where 0 represents informative content and 1 represents emotional content.

[0034] Display mode: The value is between 0 and 1, 0 means static display, and 1 means interactive display.

[0035] Calculate the clothing style coefficient for the input , are the calculation weights of clothing type score, brand style score and display intention score, respectively. They are the clothing type score, brand style score, and display intention score corresponding to the input clothing style.

[0036] The animation generation module is connected to the action matching module and generates corresponding 3D clothing catwalk animation based on the virtual catwalk actions and virtual catwalk postures.

[0037] Use 3D modeling software to create a virtual character model, add skin maps, materials and textures to the character model, bind the character model to the skeleton so that the movement of the skeleton can drive the deformation of the character model, adjust the response of each part of the model to the movement of the skeleton, map the captured catwalk action data to the skeleton of the 3D character model, and adjust the character's posture according to the reference posture data to make it conform to the standard posture of the catwalk; Use 3D modeling software to create virtual clothing models, use the physics engine to simulate the dynamic behavior of clothing, including the wrinkles, swings and interaction of fabric with the character's body, bind the clothing model to the character model, adjust the clothing model and physical parameters based on the simulation results, and export the final 3D catwalk animation as a video or other suitable format. Example

[0038] See also Figure 2 As shown, the animation generation method based on the human body structure includes the following steps: Collect the three-dimensional coordinates of each joint point corresponding to each difference individual; Process the three-dimensional coordinates of each joint point of each individual to generate a standardized human joint point model; Select corresponding virtual catwalk movements and virtual catwalk postures according to the input clothing style; Generate corresponding 3D clothing catwalk animation based on virtual catwalk movements and virtual catwalk postures.

[0039] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0040] It should be understood that determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0041] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0042] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An animation generation system based on human body structure, characterized in that: include: Data acquisition module: used to collect the three-dimensional coordinates of each joint point corresponding to each difference individual; Data processing module: connected to the data acquisition module, used to process the three-dimensional coordinates of each joint point corresponding to each different individual, and generate a standardized human joint point model; Action library module: used to store various virtual catwalk actions; An action matching module, connected to the data processing module and the action library module, for selecting corresponding virtual catwalk actions and virtual catwalk postures according to the input clothing style; The animation generation module is connected to the action matching module and generates corresponding 3D clothing catwalk animation based on the virtual catwalk actions and virtual catwalk postures.

2. The human body structure-based animation generation system according to claim 1, characterized in that: The overall operation logic of the data processing module is: Select each key node of the human body from each joint, thereby obtaining the three-dimensional coordinates of each key node of the human body corresponding to each difference individual, and record it as , i is the number of each difference individual, i=1,2,...A, j is the number of each key point of the human body, j=1,2,...J; Select the left shoulder node and the right shoulder node from each key node of the human body, obtain the three-dimensional coordinates of the left shoulder node and the right shoulder node corresponding to each difference individual, and then calculate the direction vector of the left shoulder node and the right shoulder node corresponding to each difference individual , is the three-dimensional coordinate of the left shoulder node corresponding to the i-th difference individual, is the three-dimensional coordinate of the right shoulder node corresponding to the i-th difference individual; The direction vectors of the left shoulder node and the right shoulder node corresponding to each difference individual Rotate to target direction , where the specific calculation steps of the rotation matrix R are as follows: Calculate the axis of rotation , is the norm symbol; calculate the rotation angle ; The rotation matrix is ​​calculated from this , The rotation axis The cross product matrix of I is the identity matrix, are the rotation axes corresponding to the x-axis, y-axis and z-axis respectively; Rotate the key nodes of each human body corresponding to each difference individual to the target direction to obtain the rotated three-dimensional coordinates of the key nodes of each human body corresponding to each difference individual .

3. The human body structure-based animation generation system according to claim 2, characterized in that: The overall operation logic of the data processing module also includes the following steps: Obtain the rotated three-dimensional coordinates of each key node of the human body corresponding to each difference individual, calculate the mean value thereof, and obtain the mean value of the rotated three-dimensional coordinates corresponding to each key node of the human body; The standard deviation of the three-dimensional coordinates of the rotation of each key node of the human body corresponding to each difference individual is calculated synchronously to obtain the standard deviation of the three-dimensional coordinates of the rotation of each key node of the human body; The rotated three-dimensional coordinates of each key node of the human body corresponding to each difference individual are standardized using the mean and standard deviation of the rotated three-dimensional coordinates of each key node of the human body to obtain the standardized three-dimensional coordinates of each key node of the human body corresponding to each difference individual. , is the mean of the rotated three-dimensional coordinates corresponding to the jth human key node, is the standard deviation of the rotated three-dimensional coordinates corresponding to the jth human key node; The standardized three-dimensional coordinates of each key human body node corresponding to each different individual are imported into a predefined human body joint point model to generate the human body joint point model of each different individual.

4. The human body structure-based animation generation system according to claim 2, characterized in that: The specific process of generating a standardized human joint point model is as follows: Obtain the three-dimensional coordinates of each joint in the human joint point model of each difference individual, calculate the distance between each pair of adjacent nodes in each difference individual based on the three-dimensional coordinates, and then calculate the mean value of the distance between each pair of adjacent nodes; The mean distance between each pair of adjacent nodes is imported into the reference human joint point model to generate a standardized human joint point model.

5. The human body structure-based animation generation system according to claim 1, characterized in that: Select the corresponding virtual catwalk action according to the input clothing style. The specific operation logic is as follows: Extract features from the input clothing style and obtain the values ​​of each feature corresponding to the input clothing style ; And extract features from the various virtual catwalk actions stored in the action library module to obtain the values ​​of each feature corresponding to each virtual catwalk action ; The feature similarity between the input clothing style corresponding features and the features corresponding to each virtual catwalk action is calculated. , c is the number of each feature, c=1,2,...B, B is the total number of features, are the values ​​of the first and second features corresponding to the input clothing style, are the values ​​of the first feature and the second feature corresponding to the f-th virtual catwalk action, respectively, and f is the number of each virtual catwalk action; The feature similarities between the input clothing style corresponding features and the features corresponding to each virtual catwalk action are arranged in descending order, and the virtual catwalk action with the highest feature similarity is selected as the virtual catwalk action of the input clothing style.

6. The human body structure-based animation generation system according to claim 1, characterized in that: The virtual catwalk posture is divided into virtual catwalk step length and virtual catwalk speed.

7. The human body structure-based animation generation system according to claim 1, characterized in that: Select the corresponding virtual catwalk posture. The specific selection logic is: Calculate the input clothing style coefficient based on the input clothing style ; According to the standardized human joint point model, the leg length L corresponding to the standardized human joint point model is obtained; The virtual catwalk step length corresponding to the input clothing style is calculated ; Calculate the virtual catwalk speed corresponding to the input clothing style , The benchmark catwalk speed is set.

8. The human body structure-based animation generation system according to claim 1, characterized in that: Calculate the clothing style coefficient for the input , the specific calculation logic is: Obtain the clothing type score, brand style score, and display intention score corresponding to the input clothing style, and set the calculation weights of the clothing type score, brand style score, and display intention score respectively; Calculate the clothing style coefficient for the input , are the calculation weights of clothing type score, brand style score and display intention score, respectively. They are the clothing type score, brand style score, and display intention score corresponding to the input clothing style.

9. The method for generating animation based on human body structure is characterized in that: The method is implemented based on the human body structure-based animation generation system according to claims 1 to 8, and comprises the following steps: Collect the three-dimensional coordinates of each joint point corresponding to each difference individual; Process the three-dimensional coordinates of each joint point of each individual to generate a standardized human joint point model; Select corresponding virtual catwalk movements and virtual catwalk postures according to the input clothing style; Generate corresponding 3D clothing catwalk animation based on virtual catwalk movements and virtual catwalk postures.

Images