Method, system, medium and equipment for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation

Abstract

The present invention discloses a virtual classroom teaching implementation method, system, medium and equipment based on homology mapping three-dimensional shape estimation, and the method includes the following steps: S1, modeling and inputting: constructing a virtual classroom model and a teacher-student 3D model by combining U-V transformation and homology mapping geometric generation model, and pre-entering modeling information in a virtual classroom management workstation; S2, login: using a camera to collect and input 2D images of teachers and students; S3, identification: identifying the teacher-student 3D model based on the teacher-student 2D image based on the DensePose algorithm and the homology mapping geometric generation model; S4, authentication: establishing an authentication model based on the teacher-student 3D model, and performing authentication information interaction between the teacher-student 3D model and the modeling information in step S3 according to a preset authentication protocol, and when the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful. The present invention can realize a more flexible and vivid teaching process and provide a better offline teaching and learning experience.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing and artificial intelligence, and in particular, relates to a method, system, medium and equipment for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation. Background Art

[0002] Virtual classroom is one of the challenging areas in information technology, and its application includes a subsystem of the metaverse. This invention patent introduces the entry and authentication of 3D teacher, student and classroom environment models based on 3D shape estimation of homology mapping geometric generation model.

[0003] Traditional virtual classrooms are mainly presented in two-person form, and are mainly divided into the following types: voice PPT, video conference call, video live broadcast and pre-recorded video. Voice PPT products are represented by YuKeTang; pre-recorded videos include MOOC, which was very popular some time ago; video conference calls are implemented through office software such as Tencent Conference and DingTalk; video live broadcast is even more popular through various live broadcast platforms. However, no matter which type of virtual classroom is implemented, it is inevitable that the teacher cannot grasp the overall situation of the class and is in an embarrassing situation of teaching alone in the air, and the teaching and learning experience of the virtual classroom is poor. Summary of the invention

[0004] The first purpose of the invention is to overcome the shortcomings and deficiencies in the prior art, and to provide a virtual classroom teaching implementation method based on homology mapping three-dimensional shape estimation, so as to achieve a more flexible and vivid teaching process and improve the teaching and learning experience.

[0005] The second object of the present invention is to provide a virtual classroom implementation system based on homology mapping three-dimensional shape estimation.

[0006] A third object of the present invention is to provide a storage medium.

[0007] A fourth object of the present invention is to provide a computing device.

[0008] The purpose of the present invention is achieved by the following technical solution: a virtual classroom teaching implementation method based on homology mapping three-dimensional shape estimation, comprising the steps of:

[0009] S1. Modeling and inputting: constructing a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a coherent mapping geometric generation model, or using a 3D color scanner to construct a virtual classroom model and a 3D model of teachers and students, and pre-entering modeling information in a virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the coherent mapping geometric generation model includes a coherent mapping and a geometric generation model;

[0010] S2, login: use the camera to collect and input the 2D images of teachers and students;

[0011] S3, recognition: Based on the DensePose algorithm and the homology mapping geometric generation model, the teacher and student 3D models are recognized according to the teacher and student 2D images;

[0012] S4, authentication: establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to the preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

[0013] Preferably, in step S1, the virtual classroom model includes classroom props and their spatial configuration and teaching environment, the classroom props include a podium, a blackboard, a desk, a stool, doors and windows, a wall and a wall clock, and the teaching environment includes a square mode and an indoor mode.

[0014] Preferably, in step S1, the establishment of the teacher-student 3D model specifically includes the following steps:

[0015] S11. Constructing a homology mapping geometric generation model based on homology mapping and geometric generation model;

[0016] S22, using a camera to obtain a 2D image of the teacher and the student and inputting it into a coherent mapping geometry generation model;

[0017] S23, based on the DensePose algorithm, the precise information of the main joints and surface shape splines of the human body is extracted from the 2D images of the teacher and the students as the input conditions for 3D reconstruction;

[0018] S24. Generate a teacher-student 3D model corresponding to the input teacher-student 2D image based on the generative adversarial network W-GAN and in combination with the input conditions.

[0019] Preferably, step S3 specifically includes:

[0020] S41. Constructing a homology mapping geometric generation model based on homology mapping and geometric generation model;

[0021] S42, inputting the teacher-student 2D images of step S2 into the coherence mapping geometric generation model;

[0022] S43, UV transformation: Based on the DensePose algorithm, the precise information of human joints and surface shape splines is extracted from the 2D images of teachers and students as the input conditions for 3D reconstruction;

[0023] S44. Generate a teacher-student 3D model corresponding to the teacher-student 2D image of step S2 based on the generative adversarial network W-GAN and combined with the input conditions after UV transformation.

[0024] Preferably, the homology mapping geometric generation model is obtained based on the homology mapping and the geometric generation model, specifically including entering the data of each point of the teacher-student 3D model into the function of the homology mapping for several optimizations, solving the preset optimal mapping relationship, thereby realizing the geometric generation model, and the geometric generation model is used for estimating the teacher-student 3D model;

[0025] The expression of the homology mapping is: , the homology mapping is a critical point of the energy E(H) expressed in formula (1), which is obtained by minimizing the energy E(H):

[0026] , formula (1),

[0027] Among them, H represents homology mapping, S represents the three-dimensional original image domain, D represents the target domain after mapping, and the differential norm Through the measurement of S and D, dμ s is an area element on the 3D surface, and the approximate solution of E(H) is:

[0028] , formula (2),

[0029] Among them, [v1, v2] is an edge connecting two adjacent vertices v1 and v2, k [v1，v2] Calculate according to formula (3):

[0030] , formula (3),

[0031] Among them, {v0, v1, v2} and {v0, v1, v3} are two adjacent triangular faces;

[0032] According to equation (4), the energy of equation (2) can be minimized by using the Euler-Lagrange differential equation:

[0033] , formula (4),

[0034] in, is the Laplace-Beltrami operator, which evolves into the problem of solving the sparse least squares system shown in equation (5):

[0035] , formula (5).

[0036] Preferably, in step S4, the establishment of the authentication model based on the teacher-student 3D model specifically includes the following steps:

[0037] Set A, B and C as three criteria, and criteria A, B and C all include personal identity information, personal 3D shape feature information and logical relationships of teachers and students;

[0038] Authentication information is transmitted between criterion A and criterion C, and between criterion A and criterion B. Specifically, criterion A sends a declaration message to criterion B to express its intention to communicate with criterion B, and criterion B then feeds back information to criterion A.

[0039] Preferably, in step S4, the expression of the authentication protocol is:

[0040] , formula (6),

[0041] Among them, P and Q are setting criteria, Message is the position information and 3D model information of the teacher or student, and Formula (6) represents the authentication process from setting criterion P to setting criterion Q based on Message.

[0042] The virtual classroom teaching implementation system based on homology mapping three-dimensional shape estimation includes: modeling and input module, login module, recognition module and authentication module.

[0043] The modeling and input module is used to build a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a homology mapping geometric generation model, or to build a virtual classroom model and a 3D model of teachers and students by using a 3D color scanner, and to pre-enter the modeling information in the virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the homology mapping geometric generation model includes a homology mapping and a geometric generation model;

[0044] The login module is used to collect and input 2D images of teachers and students using a camera;

[0045] The recognition module is used to recognize the teacher and student 3D models according to the teacher and student 2D images based on the DensePose algorithm and the homology mapping geometric generation model;

[0046] The authentication module is used to establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to a preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

[0047] A storage medium stores a program, and when the program is executed by a processor, the virtual classroom teaching method based on homology mapping three-dimensional shape estimation is implemented.

[0048] A computing device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the above-mentioned virtual classroom teaching implementation method based on homology mapping three-dimensional shape estimation is implemented.

[0049] Compared with the prior art, the present invention has the following advantages and effects:

[0050] (1) The present invention provides a method for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation. First, a 3D model of teachers and students and a virtual classroom model are constructed. The modeling and modeling information are pre-entered using a method combining UV transformation and homology mapping geometric generation model or directly through a 3D scanner; then the 2D images of teachers and students are acquired through a camera for login and recognition; then the reconstruction from the 2D images of teachers and students to the 3D models of teachers and students is completed based on the homology mapping geometric generation model, and the authentication is completed. Among them, matching is performed with the pre-entered modeling information according to the preset authentication protocol. When the authentication is successful, the corresponding 3D models of teachers and students are added to the 3D models of teachers and students and the virtual classroom model to realize the final login of the system. Compared with the existing conventional online teaching methods, the present invention is more vivid and flexible, closer to real offline teaching, and is also suitable for remote teaching and distributed teaching under special circumstances.

[0051] (2) The algorithm of the present invention has the advantage of being a generative model with a fast calculation speed. Moreover, due to the adoption of the optimal mapping of distribution uniformity (i.e., the coherent mapping geometric generation model), the estimation effect of the 3D model is more realistic, which can improve the teaching and learning experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 The present invention is a flowchart of a method for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation.

[0053] Figure 2 This is a schematic diagram of the principle of three-dimensional reconstruction based on the DensePose algorithm and the homology mapping geometric generation model according to Example 1 of the present invention.

[0054] Figure 3 It is a schematic diagram of the authentication model based on the teacher-student 3D model in Example 1 of the present invention.

[0055] Figure 4 This is a schematic diagram of the authentication logic based on the teacher-student 3D model in Example 1 of the present invention.

[0056] Figure 5 This is a schematic diagram of classroom props and their spatial configuration in a virtual classroom model constructed by a 3D color scanner according to the present invention, wherein (a) is a schematic diagram of the composition of classroom props, and (b) is a schematic diagram of the spatial configuration of classroom props.

[0057] Figure 6 This is a schematic diagram of the teacher-student 3D model established by the present invention through a 3D color scanner.

[0058] Figure 7 A schematic diagram of a square mode of a teaching environment of a virtual classroom model constructed by a 3D color scanner according to the present invention.

[0059] Figure 8 It is a schematic diagram of the principle of the geometric generation model of the present invention. DETAILED DESCRIPTION

[0060] The present invention will be further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0061] Example 1

[0062] like Figure 1 The figure shows a flow chart of a method for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation, including the following steps:

[0063] S1. Modeling and inputting: constructing a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a coherent mapping geometric generation model, or using a 3D color scanner to construct a virtual classroom model and a 3D model of teachers and students, and pre-entering modeling information in a virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the coherent mapping geometric generation model includes a coherent mapping and a geometric generation model;

[0064] S2, login: use the camera to collect and input the 2D images of teachers and students;

[0065] S3, recognition: Based on the DensePose algorithm and the homology mapping geometric generation model, the teacher and student 3D models are recognized according to the teacher and student 2D images;

[0066] S4, authentication: establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to the preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

[0067] Specifically, in order to realize virtual classroom teaching, it is necessary to establish a three-dimensional model of teachers, students and classroom environment. In this embodiment, the virtual classroom model and the 3D model of teachers and students can be input in two ways: one is based on the harmonic mapping geometric generative model (HMGGM) method. The other is to use a 3D color scanner.

[0068] In method 1, the UV transformation includes the DensePose algorithm, and the homology mapping geometry generative model includes the 3D human body modeling method SMPL, the harmonic maps (HM) and the geometric generative model (GMM). The present invention adopts 3D reconstruction based on the combination of the DensePose algorithm, the SMPL algorithm and the homology mapping geometry generative model (HMGMM). Among them, the 3D human body modeling method SMPL is the abbreviation of Skined Multi-Person Linear Model, which is a commonly used model for 3D human body modeling.

[0069] The 3D color scanner in method 2 may be specifically an EinScan-H 3D color three-dimensional scanner. In step S1, the virtual classroom model includes classroom props and their spatial configuration and teaching environment. The classroom props include a podium, a blackboard, a desk, a stool, doors and windows, a wall and a wall clock. The teaching environment includes a square mode and an indoor mode. Figure 5 The schematic diagram of classroom props and their spatial configuration in the virtual classroom created by 3D color scanner is shown in FIG. Figure 6 The schematic diagram of the 3D model of teachers and students established by 3D color scanner is shown in FIG. Figure 7 Shown is a schematic diagram of the teaching environment of a virtual classroom created using a 3D color scanner.

[0070] The login step uses a camera to capture and input 2D images of teachers and students, and then performs three-dimensional reconstruction and recognition based on the DensePose algorithm and the coherent mapping geometric generation model. Finally, the pre-entered modeling information is matched with the 3D model of teachers and students obtained by three-dimensional reconstruction through the authentication model. If the match is successful, login is achieved, and the corresponding 3D model of the teacher / student appears in the virtual classroom model. Compared with the existing conventional online teaching methods, the present invention is more vivid and flexible, closer to real offline teaching, and is also suitable for remote teaching and distributed teaching under special circumstances.

[0071] In step S1, the establishment of the teacher-student 3D model specifically includes the following steps:

[0072] S11. Constructing a homology mapping geometric generation model based on homology mapping and geometric generation model;

[0073] S22, using a camera to obtain a 2D image of the teacher and the student and inputting it into a coherent mapping geometry generation model;

[0074] S23, based on the DensePose algorithm, the precise information of the main joints and surface shape splines of the human body is extracted from the 2D images of the teacher and the students as the input conditions for 3D reconstruction;

[0075] S24. Generate a teacher-student 3D model corresponding to the input teacher-student 2D image based on the generative adversarial network W-GAN and in combination with the input conditions.

[0076] Step S3 specifically includes:

[0077] S41. Constructing a homology mapping geometric generation model based on homology mapping and geometric generation model;

[0078] S42, inputting the teacher-student 2D images of step S2 into the coherence mapping geometric generation model;

[0079] S43, UV transformation: Based on the DensePose algorithm, the precise information of human joints and surface shape splines is extracted from the 2D images of teachers and students as the input conditions for 3D reconstruction;

[0080] S44. Generate a teacher-student 3D model corresponding to the teacher-student 2D image of step S2 based on the generative adversarial network W-GAN and combined with the input conditions after UV transformation.

[0081] Specifically, before the virtual classroom can identify the identity of teachers and students, the present invention proposes a three-dimensional reconstruction based on the homology mapping geometric generation model HMGGM from simple 2D to 3D image conversion, such as Figure 2 It is a schematic diagram showing the principle of three-dimensional reconstruction based on the coherent mapping geometric generation model of the present invention.

[0082] Among them, UV transformation uses the illumination method of computer vision. The specific computer algorithm is DensePose, which calculates the 3D surface shape from the 2D image. Figure 2 The HMGGM encoder, HMGGM decoder and discriminator output verification together constitute the Wasserstein Generative Adversarial Networks (W-GAN). W-GAN has been trained with a large number of 2D-3D model data sets and is capable of calculating the shape of 2D images to 3D surfaces.

[0083] The homology mapping geometric generation model is obtained based on the homology mapping and the geometric generation model, specifically including entering the data of each point of the teacher-student 3D model into the function of the homology mapping for several optimizations, solving the preset optimal mapping relationship, thereby realizing the geometric generation model, and the geometric generation model is used for estimating the teacher-student 3D model;

[0084] The expression of the homology mapping is: , the homology mapping is a critical point of the energy E(H) expressed in formula (1), which is obtained by minimizing the energy E(H):

[0085] , formula (1),

[0086] Where H represents the homology mapping, S represents the three-dimensional original image domain, and D represents the target domain after mapping. The entire formula (1) represents the homology mapping from S to D. The differential norm Through the measurement of S and D, dμ s is an area element on the 3D surface, and the approximate solution of E(H) is:

[0087] , formula (2),

[0088] Among them, [v1, v2] is an edge connecting two adjacent vertices v1 and v2, k [v1，v2] Calculate according to formula (3):

[0089] , formula (3),

[0090] Among them, {v0, v1, v2} and {v0, v1, v3} are two adjacent triangular faces;

[0091] According to equation (4), the energy of equation (2) can be minimized by using the Euler-Lagrange differential equation:

[0092] , formula (4),

[0093] in, is the Laplace-Beltrami operator, which evolves into the problem of solving the sparse least squares system shown in equation (5):

[0094] , formula (5).

[0095] Specifically, homology mapping is a type of comformal mapping. In this embodiment, Figure 8 The figure shows the principle diagram of the geometric generative model. The geometric generative model GMM is used to estimate the teacher and student 3D models. Its significance is that the relative relationship of geometric elements remains unchanged during the transformation from 2D to 3D or from 3D to 2D, which is also called angle-preserving transformation. Figure 8 In it, x is the image space, z is the feature space after dimensionality reduction, and z is statistically uniformly distributed.

[0096] In principle, the geometric generation model can be realized by entering each point of the 3D model and then performing multiple optimizations through the function of homological mapping of equations (1) to (5) or solving the optimal mapping relationship through multiple vertices of the geometric mesh. In this embodiment, the following approximate two-step numerical algorithm is used:

[0097] (1) The data in the image space x is mapped to the feature space z through the artificial neural network W-GAN, where the dimension can be analyzed;

[0098] (2) Perform probability density conversion. The image space data conforms to a probability distribution. This original probability distribution can be converted into a specified probability distribution through mapping. This second step uses a geometric method.

[0099] Therefore, the advantage of the algorithm of the present invention is that it is a kind of generative model with a fast calculation speed, and because it adopts the optimal mapping of distribution uniformity (i.e., the coherent mapping geometric generation model), the estimation effect of the 3D model is more realistic, which can improve the teaching and learning experience.

[0100] In step S4, the establishment of the authentication model based on the teacher-student 3D model specifically includes the following steps:

[0101] Set A, B and C as three criteria, and criteria A, B and C all include personal identity information, personal 3D shape feature information and logical relationships of teachers and students;

[0102] Authentication information is transmitted between criterion A and criterion C, and between criterion A and criterion B. Specifically, criterion A sends a declaration message to criterion B to express its intention to communicate with criterion B, and criterion B then feeds back information to criterion A.

[0103] Specifically, the authentication between the criteria is completed with the help of a mobile phone APP or a computer program. Figure 3 FIG. 1 is a schematic diagram of the authentication model based on the teacher-student 3D model of the present invention. Authentication information is sent or received between criteria. Figure 4 The figure shows the authentication logic diagram based on the teacher-student 3D model of the present invention. New user registration requires the entry of corresponding 3D model information. When the user logs in, the 2D teacher image or student image is used to reconstruct into a 3D model for authentication, and is matched with the teacher-student 3D model pre-entered in the virtual classroom management workstation to achieve login.

[0104] In step S4, the expression of the authentication protocol is:

[0105] , formula (6),

[0106] Among them, P and Q are setting criteria, Message is the position information and 3D model information of the teacher or student, and Formula (6) represents the authentication process from setting criterion P to setting criterion Q based on Message.

[0107] Specifically, the common authentication logic can express the credit of a trust pair, which involves an authentication protocol. In this embodiment, this type of authentication protocol is represented by formula (6).

[0108] Example 2

[0109] The virtual classroom teaching implementation system based on homology mapping three-dimensional shape estimation includes: modeling and input module, login module, recognition module and authentication module.

[0110] The modeling and input module is used to build a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a homology mapping geometric generation model, or to build a virtual classroom model and a 3D model of teachers and students by using a 3D color scanner, and to pre-enter the modeling information in the virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the homology mapping geometric generation model includes a homology mapping and a geometric generation model;

[0111] The login module is used to collect and input 2D images of teachers and students using a camera;

[0112] The recognition module is used to recognize the teacher and student 3D models according to the teacher and student 2D images based on the DensePose algorithm and the homology mapping geometric generation model;

[0113] The authentication module is used to establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to a preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

[0114] Example 3

[0115] A storage medium stores a program, and when the program is executed by a processor, the virtual classroom teaching method based on homology mapping three-dimensional shape estimation described in Example 1 is implemented as follows:

[0116] S1. Modeling and inputting: constructing a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a coherent mapping geometric generation model, or using a 3D color scanner to construct a virtual classroom model and a 3D model of teachers and students, and pre-entering modeling information in a virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the coherent mapping geometric generation model includes a coherent mapping and a geometric generation model;

[0117] S2, login: use the camera to collect and input the 2D images of teachers and students;

[0118] S3, recognition: Based on the DensePose algorithm and the homology mapping geometric generation model, the teacher and student 3D models are recognized according to the teacher and student 2D images;

[0119] S4, authentication: establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to the preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

[0120] In the above process, the specific processing process is as described in Example 1 and will not be repeated here.

[0121] In this embodiment, the storage medium may be a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, a USB flash drive, a mobile hard disk, or the like.

[0122] Example 4

[0123] A computing device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the virtual classroom teaching implementation method based on homology mapping three-dimensional shape estimation described in Example 1 is implemented as follows:

[0124] S1. Modeling and inputting: constructing a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a coherent mapping geometric generation model, or using a 3D color scanner to construct a virtual classroom model and a 3D model of teachers and students, and pre-entering modeling information in a virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the coherent mapping geometric generation model includes a coherent mapping and a geometric generation model;

[0125] S2, login: use the camera to collect and input the 2D images of teachers and students;

[0126] S3, recognition: Based on the DensePose algorithm and the homology mapping geometric generation model, the teacher and student 3D models are recognized according to the teacher and student 2D images;

[0127] S4, authentication: establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to the preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

[0128] In the above process, the specific processing process is as described in Example 1 and will not be repeated here.

[0129] In this embodiment, the computing device may be a terminal device such as a desktop computer, a laptop computer, a PDA handheld terminal, a tablet computer, etc.

[0130] The above embodiments are preferred implementations of the present invention and are not intended to limit the present invention. Any other changes or other equivalent replacement methods that do not deviate from the technical solutions of the present invention are included in the protection scope of the present invention.

Claims

1. A virtual classroom teaching implementation method based on homology mapping three-dimensional shape estimation, characterized in that: Includes steps: S1. Modeling and inputting: constructing a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a coherent mapping geometric generation model, or using a 3D color scanner to construct a virtual classroom model and a 3D model of teachers and students, and pre-entering modeling information in a virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the coherent mapping geometric generation model includes a coherent mapping and a geometric generation model; The homology mapping geometric generation model is obtained based on the homology mapping and the geometric generation model, specifically including entering the data of each point of the teacher-student 3D model into the function of the homology mapping for several optimizations, solving the preset optimal mapping relationship, thereby realizing the geometric generation model, and the geometric generation model is used for estimating the teacher-student 3D model; The expression of the homology mapping is: , the homology mapping is a critical point of the energy E(H) expressed in formula (1), which is obtained by minimizing the energy E(H): , formula (1), Among them, H represents homology mapping, S represents the three-dimensional original image domain, D represents the target domain after mapping, and the differential norm Through the measurement of S and D, dμ s is an area element on the 3D surface, and the approximate solution of E(H) is: , formula (2), Among them, [v1, v2] is an edge connecting two adjacent vertices v1 and v2, k [v1，v2] Calculate according to formula (3): , formula (3), Among them, {v0, v1, v2} and {v0, v1, v3} are two adjacent triangular faces; According to equation (4), the energy of equation (2) can be minimized by using the Euler-Lagrange differential equation: , formula (4), in, is the Laplace-Beltrami operator, which evolves into the problem of solving the sparse least squares system shown in equation (5): , formula (5); S2, login: use the camera to collect and input the 2D images of teachers and students; S3, recognition: Based on the DensePose algorithm and the homology mapping geometric generation model, the teacher and student 3D models are recognized according to the teacher and student 2D images; Step S3 specifically includes: S41. Constructing a homology mapping geometric generation model based on homology mapping and geometric generation model; S42, inputting the teacher-student 2D images of step S2 into the coherence mapping geometric generation model; S43, UV transformation: Based on the DensePose algorithm, the precise information of human joints and surface shape splines is extracted from the 2D images of teachers and students as the input conditions for 3D reconstruction; S44, generating a teacher-student 3D model corresponding to the teacher-student 2D image of step S2 based on the generative adversarial network W-GAN and combined with the input conditions after UV transformation; S4, authentication: establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to the preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

2. The method for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation according to claim 1 is characterized in that: In step S1, the virtual classroom model includes classroom props and their spatial configuration and teaching environment. The classroom props include a podium, a blackboard, a desk, a stool, doors and windows, a wall and a wall clock. The teaching environment includes a square mode and an indoor mode.

3. The method for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation according to claim 1, characterized in that: In step S1, the establishment of the teacher-student 3D model specifically includes the following steps: S11. Constructing a homology mapping geometric generation model based on homology mapping and geometric generation model; S22, using a camera to obtain a 2D image of the teacher and the student and inputting it into a coherent mapping geometry generation model; S23, based on the DensePose algorithm, the precise information of the main joints and surface shape splines of the human body is extracted from the 2D images of the teacher and the students as the input conditions for 3D reconstruction; S24. Generate a teacher-student 3D model corresponding to the input teacher-student 2D image based on the generative adversarial network W-GAN and in combination with the input conditions.

4. The method for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation according to claim 1, characterized in that: In step S4, the establishment of the authentication model based on the teacher-student 3D model specifically includes the following steps: Set A, B and C as three criteria, and criteria A, B and C all include personal identity information, personal 3D shape feature information and logical relationships of teachers and students; Authentication information is transmitted between criterion A and criterion C, and between criterion A and criterion B. Specifically, criterion A sends a declaration message to criterion B to express its intention to communicate with criterion B, and criterion B then feeds back information to criterion A.

5. The method for implementing virtual classroom teaching based on homology mapping three-dimensional shape estimation according to claim 4 is characterized in that: In step S4, the expression of the authentication protocol is: , formula (6), Among them, P and Q are setting criteria, Message is the position information and 3D model information of the teacher or student, and Formula (6) represents the authentication process from setting criterion P to setting criterion Q based on Message.

6. A virtual classroom teaching implementation system based on homology mapping three-dimensional shape estimation, characterized in that: The method applied to claim 1 comprises: a modeling and input module, a login module, an identification module and an authentication module, The modeling and input module is used to build a virtual classroom model and a 3D model of teachers and students by combining UV transformation and a homology mapping geometric generation model, or to build a virtual classroom model and a 3D model of teachers and students by using a 3D color scanner, and to pre-enter the modeling information in the virtual classroom management workstation; the UV transformation includes a DensePose algorithm, and the homology mapping geometric generation model includes a homology mapping and a geometric generation model; The login module is used to collect and input 2D images of teachers and students using a camera; The recognition module is used to recognize the teacher and student 3D models according to the teacher and student 2D images based on the DensePose algorithm and the homology mapping geometric generation model; The authentication module is used to establish an authentication model based on the teacher-student 3D model, and perform authentication information interaction on the teacher-student 3D model and modeling information in step S3 according to a preset authentication protocol. When the authentication is successful, the virtual classroom model adds the corresponding teacher-student 3D model, indicating that the login system is successful.

7. A storage medium storing a program, characterized in that: When the program is executed by a processor, the virtual classroom teaching implementation method based on homology mapping three-dimensional shape estimation described in any one of claims 1 to 5 is implemented.

8. A computing device, characterized in that It comprises a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the virtual classroom teaching implementation method based on homology mapping three-dimensional shape estimation described in any one of claims 1 to 5 is implemented.

Images