Method, device and equipment for copyright identification of three-dimensional mesh model and storage medium

Abstract

The application discloses a copyright identification method, device and equipment of a three-dimensional grid model and a storage medium. The copyright identification method of the three-dimensional grid model comprises the following steps: obtaining a three-dimensional network model to be identified; segmenting the three-dimensional grid model based on a preset segmentation method to obtain a plurality of blocks; based on a first vertex of each block in the plurality of blocks, a plurality of bounding boxes are constructed in each block, the bounding box is connected with all vertices adjacent to the base point to form a first vertex as a base point, and the first vertex is a vertex satisfying a preset base point condition; based on a preset algorithm, a characteristic value corresponding to a target angle is calculated, the target angle is an angle in a triangular facet angle of each bounding box in the plurality of bounding boxes, and the angle satisfies a preset condition; based on a plurality of characteristic values, a zero watermark is determined, and the zero watermark is used to indicate the copyright of the three-dimensional network model. The copyright identification method of the three-dimensional grid model provided by the application improves the available feature data amount of the three-dimensional network model and enhances the robustness of the watermark.

Description

Technical Field

[0001] This application relates to the field of communication technology, specifically to a method, apparatus, device, and storage medium for copyright identification of a three-dimensional mesh model. Background Technology

[0002] Existing digital watermarking for 3D mesh models involves directly embedding the watermark into the 3D mesh model. This approach uses the geometric information of the 3D mesh as the carrier of the watermark information, and then embeds the watermark by modifying the spatial or frequency domain information of certain 3D mesh models.

[0003] The trade-off between transparency and robustness in digital watermarking has always been a challenge in 3D mesh model watermarking technology. To address this challenge, zero-watermarking emerged. Zero-watermarking offers strong robustness and maintains transparency without altering the model's data. However, extracting feature information becomes a new challenge after adopting zero-watermarking. Although 3D mesh models are spatially based and contain far more data than 1D or 2D models, extracting the feature information that represents them is extremely difficult, leading to difficulties in copyright identification of 3D mesh models. Summary of the Invention

[0004] The purpose of this application is to provide a method, apparatus, device, and storage medium for copyright identification of three-dimensional mesh models, which can solve the technical problems of difficulty in extracting feature information and small amount of feature information data in the prior art.

[0005] The technical solution of this application is as follows:

[0006] Firstly, a copyright identification method for a 3D mesh model is provided, comprising: acquiring a 3D mesh model to be identified; dividing the 3D mesh model into multiple blocks based on a preset segmentation method; constructing multiple bounding boxes within each of the multiple blocks based on the first vertex of each block, wherein each bounding box is formed by connecting all vertices adjacent to the first vertex as the base point, and the first vertex is a vertex that satisfies a preset base point condition; calculating the feature value corresponding to a target angle based on a preset algorithm, wherein the target angle is the angle among the triangular facet angles of each bounding box that satisfies a preset condition; and determining a zero watermark based on the multiple feature values, wherein the zero watermark is used to indicate the copyright of the 3D mesh model.

[0007] In some embodiments, the preset segmentation method adopts the octree space segmentation method.

[0008] In some embodiments, the 3D network model is segmented based on a preset segmentation method, including:

[0009] The 3D network model is placed inside the target root cube for segmentation. The target root cube is the smallest cube that encloses the 3D network model.

[0010] The segmentation stops when the side length of the segmented sub-cubes reaches a preset threshold.

[0011] In some embodiments, the included angle of the triangular facets is the included angle between the normal vectors of adjacent faces.

[0012] In some embodiments, the preset base point condition is a vertex with a number of adjacent vertices not less than 3.

[0013] In some embodiments, before constructing multiple bounding boxes within each of the multiple blocks based on a first vertex of each block, the method further includes:

[0014] Based on the topological relationships of multiple vertices in a 3D mesh model, determine the number of adjacent vertices for each vertex.

[0015] In some embodiments, determining a zero watermark based on multiple feature values ​​includes:

[0016] Divide multiple feature values ​​into multiple groups according to their numerical values;

[0017] Determine the number of feature values ​​contained in each of the multiple groups;

[0018] Calculate the average number of feature values ​​in each of the multiple groups;

[0019] Set the element value of the group whose number of feature values ​​is greater than the average value to 1;

[0020] Set the element value of the group whose number of feature values ​​is no greater than the average value to 0;

[0021] The element values ​​corresponding to each group in multiple groups are arranged according to the initial sorting of the multiple groups to form feature information;

[0022] Convert the feature information into a Boolean matrix;

[0023] Zero watermarking of a 3D network model is determined based on Boolean matrices.

[0024] In some embodiments, before dividing the 3D mesh model into multiple blocks based on a preset segmentation method, the method further includes:

[0025] The 3D network model is normalized based on a preset processing method.

[0026] The 3D mesh model is divided into multiple blocks based on a pre-defined segmentation method, specifically including:

[0027] The normalized 3D mesh model is divided into multiple blocks based on a preset segmentation method.

[0028] In some embodiments, the three-dimensional network model is normalized based on a preset processing method, including:

[0029] Based on multiple normals of the 3D mesh model, determine the histogram of normal distribution of the 3D mesh model;

[0030] The classification results of 3D mesh models are obtained based on the normal distribution histogram.

[0031] Based on a preset method, the principal axes of the 3D mesh model are calculated;

[0032] Based on the classification results, the principal axes of the 3D mesh model are corrected.

[0033] In some embodiments, the method further includes: registering the zero watermark to an intellectual property protection (IPR) database.

[0034] Secondly, a copyright recognition device for a three-dimensional mesh model is provided, comprising:

[0035] The model acquisition module is used to acquire the 3D network model to be identified.

[0036] The model segmentation module is used to segment a 3D mesh model into multiple blocks based on a preset segmentation method;

[0037] The bounding box construction module is used to construct multiple bounding boxes within each of multiple blocks based on the first vertex of each block. Each bounding box is formed by connecting all vertices adjacent to the first vertex, which is a vertex that satisfies a preset base point condition.

[0038] The data processing module is used to calculate the feature value corresponding to the target angle based on a preset algorithm. The target angle is the angle among the triangular facets of each of the multiple bounding boxes that meets the preset conditions.

[0039] The zero-watermark recognition module is used to determine the zero watermark based on multiple feature values. The zero watermark is used to indicate the copyright of the 3D network model.

[0040] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the steps of the copyright identification method for the three-dimensional mesh model as described in the first aspect.

[0041] Fourthly, embodiments of this application provide a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the copyright identification method for a three-dimensional mesh model as described in the first aspect.

[0042] The technical solutions provided by the embodiments of this application bring at least the following beneficial effects:

[0043] The copyright identification method for 3D mesh models provided in this application embodiment recursively divides the 3D model into multiple blocks based on a preset segmentation method, effectively increasing the amount of usable 3D mesh model data and enhancing the robustness of the watermark. Based on each fixed point of each block, multiple bounding boxes are constructed within each block. Using a preset algorithm, feature values ​​corresponding to target angles are calculated. The target angle is the angle among the triangular facet angles of each bounding box that satisfies preset conditions. Based on multiple feature values, a zero watermark is determined. Determining feature information by taking the triangular facet angles within the bounding boxes maximizes the reflection of the shape of the 3D model.

[0044] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0045] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application, and do not constitute an undue limitation of this application.

[0046] Figure 1 This is a flowchart illustrating a copyright identification method for a three-dimensional mesh model provided in an embodiment of this application;

[0047] Figure 2 This is a schematic diagram of the structure of a bounding box provided in an embodiment of this application. Figure 1 ;

[0048] Figure 3 This is a schematic diagram of the structure of a bounding box provided in an embodiment of this application. Figure 2 ;

[0049] Figure 4 This is a schematic diagram of the expanded structure of a block provided in an embodiment of this application;

[0050] Figure 5 This is a schematic diagram of a feature-prominent part in a three-dimensional mesh model provided in an embodiment of this application;

[0051] Figure 6 This is a schematic diagram of the normalization process for the three-dimensional network model provided in the embodiments of this application;

[0052] Figure 7 This is a schematic diagram of a Bunny segmentation block with a recursion depth parameter of 30 provided in an embodiment of this application;

[0053] Figure 8This is a schematic diagram of a Bunny segmentation block with a recursion depth parameter of 100 provided in an embodiment of this application;

[0054] Figure 9 This is a schematic diagram of a Bunny segmentation block with a recursion depth parameter of 1000 provided in an embodiment of this application;

[0055] Figure 10 This is a schematic diagram of feature value grouping in an embodiment of this application;

[0056] Figure 11 This is a schematic diagram of the feature information in the embodiments of this application;

[0057] Figure 12 This is a schematic diagram of the second-order wavelet decomposition of matrix M in an embodiment of this application;

[0058] Figure 13 This is a schematic diagram of the structure of a copyright recognition device for a three-dimensional mesh model provided in an embodiment of this application;

[0059] Figure 14 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0060] To enable those skilled in the art to better understand the technical solutions of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0061] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples consistent with some aspects of this application as detailed in the appended claims.

[0062] As can be seen from the background technology, the feature information of the existing three-dimensional mesh model is difficult to extract, the amount of feature information data is small, and the copyright recognition efficiency is low.

[0063] Specifically, three-dimensional mesh models are based on space, and compared with other multimedia data types, their data volume is much larger than that of one-dimensional and two-dimensional models. However, it is extremely difficult to extract the feature information that can express the three-dimensional mesh model.

[0064] Based on the above findings, this application provides a method, apparatus, device, and storage medium for copyright identification of 3D mesh models. The 3D model is recursively segmented using a preset segmentation method to obtain multiple blocks, effectively increasing the amount of usable 3D mesh model data and enhancing the robustness of the watermark. Multiple bounding boxes are constructed within each block based on each fixed point of each of the multiple blocks. Based on a preset algorithm, feature values ​​corresponding to target angles are calculated. The target angle is the angle among the triangular facet angles of each bounding box that satisfies preset conditions. Based on multiple feature values, a zero watermark is determined. Using the triangular facet angles within the bounding boxes to determine feature information maximizes the reflection of the shape of the 3D model.

[0065] The copyright identification method for the three-dimensional mesh model provided in this application embodiment will be described in detail below with reference to the accompanying drawings.

[0066] Figure 1 This document illustrates a flowchart of a copyright identification method for a 3D mesh model provided in an embodiment of this application. Figure 1 As shown, the method may include steps S110-S150.

[0067] Step S110: Obtain the 3D network model to be identified.

[0068] A 3D model is a polygonal representation of an object, typically displayed using a computer or other video equipment. The displayed object can be a real-world entity or a fictional object. Anything that exists in the physical world can be represented using a 3D model.

[0069] 3D models are often generated using specialized software such as 3D modeling tools, but they can also be generated using other methods. As a set of points and other information, 3D models can be generated manually or according to certain algorithms. 3D models have advantages such as being intuitive, vivid, detailed, and complete, and have gradually become an important component of GIS data representation and a crucial element in the construction of smart cities and digital earth. They have also shown great application potential in many fields such as urban planning, wireless communication, disaster management, cadastral management, tourism, transportation, and environmental simulation. With the continuous expansion of the application fields of 3D models and the ease of storage, transmission, and copying, the problems of illegal copying, modification, dissemination, and misappropriation of 3D models are becoming increasingly serious, causing huge economic losses to 3D model producers. The purpose of obtaining the 3D network model in step S110 is to identify the copyright of the 3D network model in subsequent steps.

[0070] The three-dimensional network model obtained in step S110 can be the three-dimensional design source file information of the three-dimensional network model, or it can be video information, image information, etc. that can show the characteristics of the three-dimensional network model.

[0071] The 3D network model can be a 3D network model that employs digital watermarking technology for anti-counterfeiting. Specifically, this digital watermarking technology can be zero-watermark; the following description will focus on zero-watermarking.

[0072] Step S120: The three-dimensional mesh model is divided into multiple blocks based on a preset segmentation method.

[0073] The 3D network model has limited available feature data. To obtain more feature data for copyright identification, the 3D network model was segmented.

[0074] A block refers to a component of a 3D network model after it has been divided. Blocks can include large blocks and small blocks. Specifically, a large block can contain multiple small blocks, and all the large and small blocks together constitute the 3D network model.

[0075] There are several preset segmentation methods, such as octree space segmentation and quadtree segmentation. During the segmentation process, a recursive depth parameter can be set, and segmentation stops when the segmented 3D network model meets the set recursive parameter requirements.

[0076] Step S130: Based on the first vertex of each of the multiple blocks, construct multiple bounding boxes within each of the multiple blocks. Each bounding box is formed by connecting all vertices adjacent to the first vertex, with the first vertex being the base point. The first vertex is a vertex that satisfies the preset base point condition.

[0077] As an example, the preset base point condition can be to set a vertex with at least 3 adjacent vertices as the first vertex, i.e., the base point. Therefore, step S130 can specifically include steps S131-S132.

[0078] Step S131: Based on the topological relationship of multiple vertices in the 3D mesh model, determine the number of adjacent vertices of each vertex among the multiple vertices.

[0079] Based on the topological structure of the 3D mesh model, the adjacent points of each vertex can be found, and these adjacent points form several triangular patches. For example... Figure 2 As shown, this structure is a very simple bounding box, where the vertex that serves as the base point connects to three adjacent vertices.

[0080] Step S132: Based on the first vertex of each of the multiple blocks, construct multiple bounding boxes within each of the multiple blocks.

[0081] A bounding box is formed by connecting all vertices adjacent to the first vertex, with the first vertex as the base point. The simplest bounding box is one where the base point is connected to three adjacent vertices, such as... Figure 2 As shown. Furthermore, as... Figure 3 The image shows a bounding box where the base point connects to six adjacent vertices.

[0082] Two adjacent bounding boxes constructed will share two or more triangular faces, such as... Figure 4 As shown, this diagram is part of a block's unfolded diagram, where triangles a and b appear not only in the bounding box of vertex V1 but also in the bounding box of vertex V2.

[0083] Step S140: Based on a preset algorithm, calculate the feature value corresponding to the target angle. The target angle is the angle among the triangular facets of each bounding box that satisfies the preset conditions.

[0084] As an example, the included angle of a triangular facet is the angle between the normal vectors of adjacent faces. Please refer to [reference needed]. Figure 3 In this bounding box, the base point V connects to six adjacent vertices, forming six triangular faces t1, t2, t3, t4, t5, and t6. n1 and n2 are the normal vectors belonging to t1 and t2, respectively. The angle θ between n1 and n2 is the target angle.

[0085] The eigenvalues ​​of the target angle can be calculated in various ways, such as the sine or cosine of the angle, or simply the angle value itself. In the following explanation, the eigenvalues ​​will be illustrated using the cosine value as an example.

[0086] In step S140, the included angles that meet the preset conditions can be a threshold value, i.e., an included angle threshold α. For example, included angles less than 90 degrees are retained, and angles θ with included angle values ​​lower than α are retained, i.e., the corresponding cosθ values ​​are retained. Alternatively, it can be an angle range, such as included angles from 15 degrees to 70 degrees. Choosing acute angles here is more conducive to showcasing more features of the 3D network model and is easier to identify copyright information compared to obtuse angles. Therefore, by using this preset condition to eliminate the included angles of surface patches in gently sloping areas of the model's surface and retaining the included angle values ​​of surface patches with drastic local changes, it is more beneficial to identify the copyright of the 3D network model.

[0087] The angle between local spatial surfaces in the model does not change significantly; otherwise, the model would be distorted. Therefore, this property is very stable and suitable for constructing eigenvalues. For ease of calculation, this scheme uses the angle between the normal vectors of adjacent surfaces to measure the angle between adjacent surfaces. Although they are not the same concept, the angle between surfaces is actually the angle between the normal vectors of the surfaces.

[0088] Within an octree block, several triangular faces are connected by vertices, such as... Figure 3 Given that t1, t2, t3, t4, t5, and t6 are all within the same bounding box, find the angles between all pairs of faces within this bounding box. Let A(x1, y1, z1), B(x2, y2, z2), and C(x3, y3, z3) be the three vertices of the triangle face, and let the normal vector of the face be n(x...). n ,y n ,z n The cosine of the angle between the surfaces is actually the cosine of the angle between the normal vectors, and the calculation formula (1) is as follows:

[0089]

[0090] The angles between faces are related to the local details of the mesh surface. In areas lacking geometric detail, the connections between triangular faces appear relatively flat, while in areas with more prominent local details, the connections between triangular faces appear more prominent, such as... Figure 5 The area in the middle circle is where the geometric features are prominent.

[0091] In 3D mesh models, the features of localized facet regions exhibiting drastic changes are very stable. Even if the model is attacked, it is impossible to significantly alter these drastically changing triangular facet regions; otherwise, the 3D mesh model would become distorted. To reduce the computational cost and improve the uniqueness of the zero-watermark solution, this approach removes the facet angles in gently changing regions, retaining only the facet angle values ​​of those with significant local variations. Therefore, this application establishes an angle threshold α, retaining angles θ values ​​lower than α, i.e., retaining the corresponding cosθ values.

[0092] Step S150: Based on multiple feature values, determine the zero watermark, which is used to indicate the copyright of the 3D network model.

[0093] There are various methods for constructing zero watermarks, which will not be elaborated here. This application can increase the number of feature values ​​that can express a 3D mesh model. With an increased number of feature values, the available methods for constructing zero watermarks also increase accordingly.

[0094] In one example, step S150 may be followed by step S160.

[0095] Step S160: Register the zero watermark in the Intellectual Property Protection (IPR) database.

[0096] A trusted IPR (Intellectual Property Protection) database, maintained by a central authority, is responsible for verifying ownership of digital products. Zero watermarking also requires such an authority, where creators can register for zero watermarks and use the registration date in the database to avoid multiple copyright or ownership claims.

[0097] The copyright identification method for 3D mesh models provided in this application is unique in its feature extraction. Specifically, the extracted features are based on the shape characteristics of the 3D mesh model. Starting from the overall mesh model, the 3D model is sequentially segmented until the local information of the mesh model is divided into several blocks. Feature values ​​are extracted repeatedly in this manner. Since the segments consist of large blocks containing smaller blocks, extracting the feature values ​​of the 3D mesh model within the larger blocks also extracts the feature values ​​of the smaller blocks. This ensures that the local feature values ​​of the 3D model are repeatedly extracted. Furthermore, because the local details of each model are different, this not only helps improve the robustness of the zero-watermark but also strengthens its uniqueness. Secondly, by selecting, extracting, and optimizing the features of the 3D mesh model, and then using a second-order wavelet transform to process the optimized feature values, stable low-to-mid-frequency coefficients are selected to construct the zero-watermark, fundamentally guaranteeing the robustness and uniqueness of the zero-watermark.

[0098] In some embodiments, different 3D network models have different states before segmentation. Preprocessing before segmentation can make the subsequent feature value calculation process more convenient. Based on this, step S120 in the above embodiments can specifically include steps S121-S122.

[0099] Step S121: Normalize the 3D network model based on the preset processing method.

[0100] There are many preprocessing methods. The following section uses the improved principal component analysis method (NCPCA) as an example for illustration.

[0101] First, calculate the histogram of normal distribution for the 3D mesh model. Then, use the histogram to categorize the 3D mesh models. Based on the distribution characteristics of their respective normals, after calculating the principal axes of the 3D mesh models using the CPCA method, the principal axes are specifically corrected. Categorizing the 3D mesh models using histograms is crucial for calculating the principal axes using the NCPCA method. The processing steps are as follows: Figure 6 As shown.

[0102] As an example, based on a preset processing method, the three-dimensional network model is normalized, and step S121 can specifically include steps S1211-S1214.

[0103] Step S1211: Based on multiple normals of the 3D mesh model, determine the histogram of normal distribution of the 3D mesh model;

[0104] The multiple normals of a 3D network model include the normals of each face in the 3D network model.

[0105] Step S1212: Classify the 3D mesh model based on the normal distribution histogram to obtain the classification result;

[0106] Use histograms to categorize 3D mesh models.

[0107] Step S1213: Calculate the principal axes of the three-dimensional mesh model based on a preset method;

[0108] Based on the distribution characteristics of their respective normals, after calculating the principal axes of the 3D mesh model using the CPCA method, the principal axes are corrected in a targeted manner.

[0109] Step S1214: Based on the classification results, correct the principal axes of the three-dimensional mesh model.

[0110] After the three principal axes of the model are determined, the model is rotated around the centroid so that the three principal axes coincide with the z-axis, y-axis and x-axis in sequence. This completes the rotation transformation of the three-dimensional model, which is the normalization process of the three-dimensional network model.

[0111] Step S122: The normalized 3D mesh model is divided into multiple blocks based on a preset segmentation method.

[0112] There are various preset partitioning methods, such as octree partitioning and quadtree partitioning. The following text uses octree partitioning as an example.

[0113] As an example, octree segmentation of a 3D mesh model can recursively divide the 3D mesh model from the whole to the local. The smaller the recursive parameter, the more thoroughly the 3D mesh model is segmented. Based on this, step S122 can specifically include steps S1221-S1222.

[0114] Step S1221: Place the 3D network model into the target root cube for segmentation. The target root cube is the smallest cube that surrounds the 3D network model.

[0115] Set a minimum bounding size for the model and create the first root cube. Place the model elements into the cube (a cube with no child nodes). The recursive depth parameter can also be set before segmentation.

[0116] Step S1222: Stop the segmentation when the side length of the segmented sub-cubes reaches a preset threshold.

[0117] If the side length of the segmented sub-cube does not reach the set side length precision (i.e., the preset threshold), and model elements exist within the cube, then segmentation continues; if the number of model elements within the sub-cube is the same as the number of model elements within the parent cube, then segmentation stops. Segmentation stops when the side length of the sub-cube reaches the set recursion depth.

[0118] Please refer to this as an example. Figure 7-9 Bunny partitions with recursive depth parameters of 30, 100, and 1000. Figure 7 The recursion depth parameter is set to 30; Figure 8 The recursion depth parameter is set to 100; Figure 9 The recursion depth parameter is set to 1000.

[0119] Using octree properties to segment the 3D mesh model step by step makes the segments closely connected to each other. The points and faces of the bottom-level blocks are contained within the blocks of the previous level, and this process is repeated until the original model is reached. This is beneficial for extracting a large amount of 3D mesh model data.

[0120] Within each block, a bounding box is created for each vertex, connecting all vertices to the same vertex; the number of bounding boxes corresponds to the number of vertices. This method effectively represents the shape of the 3D mesh model. Then, the cosine of the angle between faces within each bounding box is calculated. A face within one bounding box will also appear in other bounding boxes, and identical cosine values ​​are extracted repeatedly. Furthermore, the cosine values ​​calculated within each octree block are also found in its parent block. This ensures the integrity and robustness of the extracted features, which is crucial for obtaining the geometric feature information of the 3D mesh model.

[0121] In some embodiments, the feature value in this embodiment is illustrated using the cosine of the target included angle θ as an example. Based on this, step S150 may specifically include steps S151-S158.

[0122] Step S151: Divide the multiple feature values ​​into multiple groups according to their numerical values.

[0123] To construct feature information from the extracted feature values, the optimized range of cosθ values ​​is P, such as... Figure 10 As shown, the range of cosθ values ​​is divided into multiple groups, i.e. Figure 10 If there are L segments in the equation, and each segment has a size of m, then P = L * m.

[0124] Step S152: Determine the number of feature values ​​contained in each of the multiple groups.

[0125] Based on the size of m, the cosθ values ​​are distributed into L small segments to count how many cosθ values ​​are in each segment.

[0126] Step S153: Calculate the average number of feature values ​​in each of the multiple groups.

[0127] The average value is obtained by dividing the total number of cosθ values ​​in multiple groups by the number of groups, which is the total number of cosθ values ​​in multiple groups divided by the value of L.

[0128] Step S154: Set the element value corresponding to the group whose number of feature values ​​is greater than the average value to 1.

[0129] Step S155: Set the element value corresponding to the group whose number of feature values ​​is not greater than the average value to 0.

[0130] Step S156: Arrange the element values ​​corresponding to each group in the multiple groups according to the initial sorting of the multiple groups to form feature information.

[0131] Three-dimensional mesh models are susceptible to various attacks, which could alter the number of cosine values ​​of the included angle within each segment. To address this issue, this application divides a sequence of length L into K groups, and averages the data within each group, setting elements greater than the average to 1 and elements less than the average to 0. This results in a sequence of either 0 or 1, while maintaining the length of L. Figure 11 As shown.

[0132] Step S157: Convert the feature information into a Boolean matrix.

[0133] Step S158: Determine the zero watermark of the 3D network model based on the Boolean matrix.

[0134] The sequence of length L is converted into a Boolean matrix M, and a second-order wavelet transform is performed on M. The low-to-mid-frequency components are then used as the zero-watermark for the 3D mesh model. Figure 12 As shown. Then register the zero watermark in the IPR database.

[0135] The extracted cosine values ​​are segmented, grouped, locally averaged to zero, and subjected to DWT2 operations in order to construct a zero watermark from the feature values. The segmentation, grouping, and locally averaged to zero operations can standardize the extracted feature data and obtain the embedding position of the zero watermark. The final DWT2 operation can enhance the robustness of the zero watermark.

[0136] Based on the same inventive concept, this application also provides a copyright recognition device for a three-dimensional mesh model.

[0137] Figure 13 This application illustrates a copyright recognition device for a three-dimensional mesh model, as shown in an embodiment. Figure 13 As shown, the copyright recognition device 200 for the 3D mesh model may include:

[0138] The model acquisition module 210 can be used to acquire the 3D network model to be identified;

[0139] The model segmentation module 220 can be used to segment the three-dimensional mesh model based on a preset segmentation method to obtain multiple blocks;

[0140] The bounding box construction module 230 can be used to construct multiple bounding boxes within each of the multiple blocks based on the first vertex of each block. The bounding box is formed by connecting all vertices adjacent to the first vertex, with the first vertex as the base point. The first vertex is a vertex that satisfies a preset base point condition.

[0141] Data processing module 240 is used to calculate the feature value corresponding to the target angle based on a preset algorithm, wherein the target angle is the angle among the triangular facets of each of the plurality of bounding boxes that satisfies the preset conditions;

[0142] The zero-watermark recognition module 250 is used to determine a zero-watermark based on multiple feature values, wherein the zero-watermark is used to indicate the copyright of the three-dimensional network model.

[0143] In some embodiments, the preset segmentation method in the model segmentation module 220 may be the octree space segmentation method.

[0144] As an example, when the model segmentation module 220 segments the 3D network model based on a preset segmentation method, the model segmentation module 220 may include segmentation units and control units.

[0145] The first segmentation unit can be used to segment the 3D network model into the target root cube, which is the smallest cube that surrounds the 3D network model.

[0146] The control unit can be used to stop the segmentation when the side length of the segmented sub-cubes reaches a preset threshold.

[0147] In some embodiments, the included angle of the triangular facets in the data processing module 240 can be the included angle of the normal vectors of adjacent faces.

[0148] In some embodiments, the preset base point condition in the bounding box construction module 230 is a vertex with a number of adjacent vertices not less than 3.

[0149] In some embodiments, the bounding box construction module 230 may include a base point determination unit and a bounding box construction unit.

[0150] The base point determination unit can be used to determine the topological relationship of multiple vertices based on a 3D mesh model, and to determine the number of adjacent vertices of each vertex among the multiple vertices.

[0151] The bounding box building unit can be used to construct multiple bounding boxes within each of multiple blocks based on the first vertex of each block, where the first vertex is a vertex with at least 3 adjacent vertices.

[0152] In some embodiments, the zero-watermark recognition module 250 may include:

[0153] The first information processing unit can be used to divide multiple feature values ​​into multiple groups according to their numerical values.

[0154] The second information processing unit can be used to determine the number of feature values ​​contained in each of the multiple groups;

[0155] The third information processing unit can be used to calculate the average number of feature values ​​in each group across multiple groups;

[0156] The first element determination unit can be used to set the element value corresponding to the group whose number of feature values ​​is greater than the average value to 1;

[0157] The second element determination unit can be used to set the element value of the group whose number of feature values ​​is not greater than the average value to 0;

[0158] The feature information generation unit can be used to arrange the element values ​​corresponding to each group in multiple groups according to the initial sorting of the multiple groups to form feature information;

[0159] The fourth information processing unit can be used to convert feature information into a Boolean matrix;

[0160] The zero-watermark determination unit can be used to determine the zero watermark of a 3D network model based on a Boolean matrix.

[0161] In some embodiments, the model segmentation module 220 may specifically include a normalization processing unit and a second segmentation unit.

[0162] The normalization processing unit can be used to normalize 3D network models based on a preset processing method.

[0163] The second segmentation unit can be used to segment the normalized 3D mesh model based on a preset segmentation method to obtain multiple blocks.

[0164] As an example, the normalization processing unit can be specifically used for:

[0165] Based on multiple normals of the 3D mesh model, determine the histogram of normal distribution of the 3D mesh model;

[0166] The classification results of 3D mesh models are obtained based on the normal distribution histogram.

[0167] Based on a preset method, the principal axes of the 3D mesh model are calculated;

[0168] Based on the classification results, the principal axes of the 3D mesh model are corrected.

[0169] In some embodiments, the copyright recognition device 200 for the three-dimensional mesh model may also include a zero-watermark registration module.

[0170] The zero-watermark registration module can be used to register zero-watermarks into the Intellectual Property Protection (IPR) database.

[0171] The copyright identification device for three-dimensional mesh models provided in this application embodiment can be used to execute the copyright identification methods for three-dimensional mesh models provided in the above method embodiments. Their implementation principles and technical effects are similar, and for the sake of brevity, they will not be described in detail here.

[0172] Based on the same inventive concept, embodiments of this application also provide an electronic device.

[0173] Figure 14 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. For example... Figure 14 As shown, the electronic device may include a processor 301 and a memory 302 storing computer programs or instructions.

[0174] Specifically, the processor 301 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of the present invention.

[0175] Memory 302 may include a large-capacity memory for data or instructions. For example, and not limitingly, memory 302 may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 502 may include removable or non-removable (or fixed) media. Where appropriate, memory 502 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 302 is a non-volatile solid-state memory. In a particular embodiment, memory 302 includes read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0176] The processor 301 reads and executes computer program instructions stored in the memory 302 to implement any of the copyright identification methods for three-dimensional mesh models in the above embodiments.

[0177] In one example, the electronic device may also include a communication interface 303 and a bus 310. For example, Figure 14 As shown, the processor 301, memory 302, and communication interface 303 are connected through bus 310 and complete communication with each other.

[0178] The communication interface 303 is mainly used to realize communication between various modules, devices, units and / or devices in the embodiments of the present invention.

[0179] Bus 310 includes hardware, software, or both, that couples components of an electronic device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 310 may include one or more buses. Although specific buses are described and illustrated in embodiments of the invention, the invention contemplates any suitable bus or interconnect.

[0180] The electronic device can execute the copyright identification method for three-dimensional mesh models in the embodiments of the present invention, thereby realizing the copyright identification method and apparatus for three-dimensional mesh models described in the above embodiments.

[0181] Furthermore, in conjunction with the copyright identification method for 3D mesh models in the above embodiments, this invention can be implemented using a readable storage medium. This readable storage medium stores program instructions; when these program instructions are executed by a processor, they implement any of the copyright identification methods for 3D mesh models described in the above embodiments.

[0182] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0183] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the required tasks. The programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0184] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0185] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. A copyright identification method for a three-dimensional mesh model, characterized in that, include: Obtain the 3D network model to be identified; The three-dimensional mesh model is divided into multiple blocks based on a preset segmentation method; Based on the first vertex of each of the plurality of blocks, a plurality of bounding boxes are constructed within each of the plurality of blocks. Each bounding box is formed by connecting all vertices adjacent to the first vertex, with the first vertex being a vertex that satisfies a preset base point condition. Based on a preset algorithm, the feature value corresponding to the target angle is calculated. The target angle is the angle among the triangular facets of each of the multiple bounding boxes that satisfies the preset conditions. Based on multiple of the aforementioned feature values, a zero watermark is determined, which is used to indicate the copyright of the three-dimensional network model; The process of determining a zero watermark based on multiple feature values ​​includes: The multiple feature values ​​are divided into multiple groups according to their numerical values; Determine the number of feature values ​​contained in each of the plurality of groups; Calculate the average number of feature values ​​in each of the multiple groups; Set the element value corresponding to the group whose number of feature values ​​is greater than the average value to 1; Set the element value corresponding to the group whose number of feature values ​​is not greater than the average value to 0; The element values ​​corresponding to each of the multiple groups are arranged according to the initial sorting of the multiple groups to form feature information; Convert the feature information into a Boolean matrix; Based on the Boolean matrix, the zero watermark of the three-dimensional network model is determined.

2. The method according to claim 1, characterized in that, The preset segmentation method adopts the octree space segmentation method.

3. The method according to claim 2, characterized in that, The segmentation of the 3D network model based on the preset segmentation method includes: The three-dimensional network model is placed inside the target root cube for segmentation, and the target root cube is the smallest cube that surrounds the three-dimensional network model; The segmentation stops when the side length of the segmented sub-cubes reaches a preset threshold.

4. The method according to claim 1, characterized in that, The included angle of the triangular facet is the included angle between the normal vectors of adjacent faces.

5. The method according to claim 1, characterized in that, The preset base point condition is a vertex with a number of adjacent vertices not less than 3.

6. The method according to claim 5, characterized in that, Before constructing multiple bounding boxes within each of the plurality of blocks based on the first vertex of each of the plurality of blocks, the method further includes: Based on the topological relationships of multiple vertices in the 3D mesh model, the number of neighboring vertices of each vertex is determined.

7. The method according to claim 1, characterized in that, Before dividing the 3D mesh model into multiple blocks based on a preset segmentation method, the method further includes: The three-dimensional network model is normalized based on a preset processing method. The process of dividing the 3D mesh model into multiple blocks based on a preset segmentation method specifically includes: The normalized 3D mesh model is divided into multiple blocks based on a preset segmentation method.

8. The method according to claim 7, characterized in that, The normalization process of the 3D network model based on the preset processing method includes: Based on the multiple normals of the three-dimensional mesh model, determine the histogram of the normal distribution of the three-dimensional mesh model; The 3D mesh model is classified based on the normal distribution histogram to obtain the classification result; The principal axes of the three-dimensional mesh model are calculated based on a preset method. Based on the classification results, the principal axes of the three-dimensional mesh model are corrected.

9. The method according to claim 1, characterized in that, The method further includes registering the zero watermark into an intellectual property protection (IPR) database.

10. A copyright recognition device for a three-dimensional mesh model, characterized in that, include: The model acquisition module is used to acquire the 3D network model to be identified. The model segmentation module is used to segment the three-dimensional mesh model into multiple blocks based on a preset segmentation method; A bounding box construction module is used to construct multiple bounding boxes within each of the multiple blocks based on the first vertex of each block. Each bounding box is formed by connecting all vertices adjacent to the first vertex, with the first vertex as the base point. The first vertex is a vertex that satisfies a preset base point condition. The data processing module is used to calculate the feature value corresponding to the target angle based on a preset algorithm. The target angle is the angle among the triangular facets of each of the multiple bounding boxes that meets the preset conditions. A zero-watermark recognition module is used to determine a zero watermark based on multiple feature values, wherein the zero watermark is used to indicate the copyright of the three-dimensional network model; The zero-watermark recognition module includes: The first information processing unit is used to divide the multiple feature values ​​into multiple groups according to their numerical values. The second information processing unit is used to determine the number of feature values ​​contained in each of the plurality of groups; The third information processing unit is used to calculate the average number of feature values ​​in each of the plurality of groups; The first element determination unit is used to set the element value corresponding to the group whose number of feature values ​​is greater than the average value to 1; The second element determination unit is used to set the element value corresponding to the group whose number of feature values ​​is not greater than the average value to 0; The feature information generation unit is used to arrange the element values ​​corresponding to each group in the plurality of groups according to the initial sorting of the plurality of groups to form feature information; The fourth information processing unit is used to convert the feature information into a Boolean matrix; The zero-watermark determination unit is used to determine the zero-watermark of the three-dimensional network model based on the Boolean matrix.

11. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the copyright identification method for a three-dimensional mesh model as described in any one of claims 1-9.

12. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the copyright identification method for a three-dimensional mesh model as described in any one of claims 1-9.

Images