CN-121981875-A - Three-dimensional model reversible information hiding method based on vertex division and smooth ordering

CN121981875ACN 121981875 ACN121981875 ACN 121981875ACN-121981875-A

Abstract

The invention discloses a three-dimensional model reversible information hiding method based on vertex division and smooth ordering, which comprises the steps of carrying out data type conversion on vertex coordinates of an original grid model by a sender to obtain a preprocessed grid model, combining secret information to be hidden, generating a secret-carrying grid model by inputting the preprocessed grid model and the secret information into a secret information embedding algorithm based on the vertex division and smooth ordering, carrying out data type conversion on the secret-carrying grid model in the same step after receiving the secret-carrying grid model by a receiver, and synchronously recovering the original grid model and complete secret information by a secret information extraction algorithm based on the vertex division and smooth ordering. The invention has good reversibility and concealment, and is suitable for the fields of copyright protection, content integrity authentication and the like of the three-dimensional model.

Inventors

WANG QINGJUN
SONG XIAOYING
ZHANG QILONG

Assignees

大连东软信息学院

Dates

Publication Date: 20260505
Application Date: 20260130

Claims (5)

1. A three-dimensional model reversible information hiding method based on vertex division and smooth ordering is characterized by comprising the following steps: s1, carrying out data type conversion on vertex coordinates of an original grid model by a sender to obtain a preprocessed grid model; s2, inputting the preprocessed grid model and secret information into a secret information embedding algorithm based on vertex division and smooth ordering for embedding processing to obtain a load-dense grid model; s3, the receiver receives the dense grid model, and performs data type conversion on the vertex coordinates of the dense grid model to obtain a preprocessed dense grid model; S4, inputting the preprocessed dense grid model into a secret information extraction algorithm based on vertex division and smooth sorting for extraction processing, and obtaining an original grid model and secret information.
2. The three-dimensional model reversible information hiding method based on vertex division and smooth ordering according to claim 1, wherein the process of embedding the secret information embedding algorithm based on vertex division and smooth ordering is as follows: S21, dividing all vertexes in the preprocessed grid model to obtain a candidate point set and a reference point set; S22, acquiring a central vertex according to the candidate point set, acquiring a ring of adjacent reference vertexes of the central vertex according to the reference point set, and constructing a plurality of embedded units according to the central vertex and the ring of adjacent reference vertexes of the central vertex; s23, calculating the predicted position coordinates of each central vertex according to the position coordinates of a ring of adjacent reference vertexes of the central vertexes, obtaining the predicted error vector of each central vertex according to the predicted position coordinates of each central vertex and the actual position coordinates of each central vertex, and combining the predicted error vectors of each central vertex into a predicted error sequence; S24, calculating the discrete Gaussian curvature of the ring of the center vertex adjacent to the reference vertex according to the position coordinate of the ring of the center vertex adjacent to the reference vertex, calculating and obtaining the local Gaussian curvature of the embedded unit according to the discrete Gaussian curvature, and determining the smoothness of the embedded unit; S25, embedding a preset auxiliary information bit stream and a load bit stream containing secret information into the ordered prediction error sequence to obtain a modified prediction error sequence; S26, adding the modified prediction error sequence and the predicted position coordinates of the central vertexes of the embedded units to obtain updated position coordinates of the vertexes of the grid model; s27, executing S21 to S26 again according to the secret grid model embedded with the initial information to finally obtain the secret grid model.
3. The three-dimensional model reversible information hiding method based on vertex division and smooth ordering according to claim 2, wherein the secret information extraction algorithm based on vertex division and smooth ordering performs the extraction processing flow as follows: s41, dividing all vertexes in the preprocessed dense grid model to obtain a candidate point set of the dense model and a reference point set of the dense model; S42, acquiring a carrier density center vertex according to a candidate point set of the carrier density model, acquiring a ring of adjacent reference vertices of the carrier density center vertex according to a reference point set of the carrier density model, and constructing a plurality of carrier density embedded units according to the carrier density center vertex and the ring of adjacent reference vertices of the carrier density center vertex; s43, calculating the predicted position coordinates of the vertexes of each dense center according to the position coordinates of a ring of adjacent reference vertexes of the vertexes of each dense center, obtaining the predicted error vector of each dense center vertex according to the predicted position coordinates of the vertexes of each dense center and the actual position coordinates of the vertexes of each dense center, and combining the predicted error vectors of the vertexes of each dense center into a dense predicted error sequence; S44, calculating the discrete Gaussian curvature of the ring of the carrier center vertex adjacent to the reference vertex according to the position coordinate of the ring of the carrier center vertex adjacent to the reference vertex, calculating and obtaining the local Gaussian curvature of the carrier embedded unit according to the discrete Gaussian curvature, and determining the smoothness of the carrier embedded unit; s45, sequentially extracting an auxiliary information bit stream and a load bit stream from the ordered load secret prediction error sequence to obtain auxiliary information, load and a recovered prediction error sequence; S46, adding the recovered prediction error sequence and the predicted position coordinates of the central vertexes of the carrier density embedded units to obtain updated position coordinates of the vertexes of the carrier density grid model; s47, executing S41 to S46 again according to the initial recovered grid model, and finally obtaining an original grid model.
4. The three-dimensional model reversible information hiding method based on vertex division and smooth sorting according to claim 2, wherein the specific steps of obtaining the prediction error vector of each center vertex are as follows: s231, calculating the predicted position coordinates of each center vertex according to a ring of adjacent reference vertexes of the center vertexes, wherein the expression is as follows: In the formula, As the central vertex Is determined by the position of the predicted position of (2); As the central vertex Is adjacent to the reference vertex set; As the central vertex Contiguous to the cardinality of the reference vertex set, an ; A ring of contiguous reference vertices that are central vertices Is a position of (2); S232, obtaining a prediction error vector of each center vertex according to the prediction position coordinates of each center vertex and the actual position coordinates of each center vertex, wherein the expression is as follows: In the formula, As the central vertex Is used for predicting the error vector of the model; As the central vertex Prediction error on the horizontal axis; As the central vertex Prediction error on the vertical axis; As the central vertex Prediction error on the depth axis; As the central vertex Is a real position coordinate vector of (a); As the central vertex Is used for predicting the position coordinate vector of the vehicle.
5. The three-dimensional model reversible information hiding method based on vertex division and smooth ordering according to claim 2, wherein the step of obtaining a local gaussian curvature of an embedded unit comprises: S241, according to any reference vertex in the embedded unit and a corresponding ring of adjacent reference vertices, obtaining an adjacent triangle of any 2 adjacent vertices in the embedded unit and the corresponding ring of adjacent reference vertices, and calculating the inner angle of the embedded unit and the corresponding Voronoi area of the adjacent triangle; s242, calculating the discrete Gaussian curvature of any reference vertex in the embedded unit according to the inner angle of the arbitrary reference vertex in the embedded unit in the adjacent triangle and the Voronoi area corresponding to the inner angle, wherein the expression is as follows: In the formula, For embedding arbitrary reference vertices in a cell Is a discrete gaussian curvature of (a); Is any reference vertex embedded in a unit in an adjacent triangle Internal corners at; for embedding arbitrary reference vertices in a cell A corresponding ring adjoins the set of reference vertices; Is any reference vertex embedded in a unit in an adjacent triangle Voronoi area of (C); S243, calculating and obtaining the local Gaussian curvature of the embedded unit according to the discrete Gaussian curvature of each reference vertex in the embedded unit, wherein the expression is as follows: In the formula, Is the local gaussian curvature of the embedded cell.

Description

Three-dimensional model reversible information hiding method based on vertex division and smooth ordering Technical Field The invention relates to the field of information security, in particular to a three-dimensional grid model reversible information hiding method based on vertex division and smooth sorting. Background With the rapid development of three-dimensional digitizing technology, three-dimensional grid models have become key carriers in the fields of digital media, computer aided design, virtual/augmented reality, cultural heritage protection, biomedicine and the like, in the applications, requirements of safe transmission, copyright authentication, content tracing, privacy information protection and the like of the models are increasingly highlighted, the three-dimensional model information hiding technology is promoted to become a research hotspot, and the technology aims to imperceptibly embed secret information (such as copyright identification, authentication watermark and additional data) into the three-dimensional models without affecting visual appearance and functional application of the secret information. The existing three-dimensional model information hiding method mainly comprises four major categories, namely a spatial domain method, a transform domain method, a compression domain method and an encryption domain method, wherein the transform domain method (such as wavelet transform and spectrum analysis) is used for converting a model into a frequency domain to carry out coefficient modification, the method has better robustness to geometric attacks (such as translation, rotation and scaling) but higher algorithm complexity, the core idea of the compression domain method is to directly embed information in a compression process or compressed code stream of a 3D model, and the encryption domain method is mainly used for carrying out information hiding on the premise of protecting the content privacy of the 3D model. In terms of space domain reversible information hiding, the mainstream technology is mainly used for directly modifying vertex coordinates, patch connection relations or vertex attributes of a model, wherein methods based on vertex coordinates are widely studied because of being direct and efficient, common strategies comprise difference-based expansion, histogram translation or prediction error expansion, however, when the methods are applied to a three-dimensional model, specific challenges are faced by the methods, namely firstly, obvious geometric distortion is easily introduced when the vertex coordinates are directly and unordered modified, smoothness and visual quality of the model surface are destroyed, secondly, how to efficiently and adaptively select a vertex subset for embedding is difficult, wherein the vertex subset is used for achieving optimal balance between high embedding capacity and low visual distortion, and finally, floating point number precision error accumulation can be generated in the embedding process or the reversibility is sensitive to vertex processing sequence, so that reversibility cannot be accurately ensured in actual calculation, and the robustness and reliability of the algorithm are affected. Disclosure of Invention The invention provides a three-dimensional model reversible information hiding method based on vertex division and smooth sorting, which aims to solve the problems of distortion and quality reduction of a model caused by unordered modification of vertex coordinates, accumulation of floating point number precision errors, precision in a reversible process, high embedding capacity and low vision. In order to achieve the above object, the technical scheme of the present invention is as follows: a three-dimensional model reversible information hiding method based on vertex division and smooth ordering comprises the following steps: s1, carrying out data type conversion on vertex coordinates of an original grid model by a sender to obtain a preprocessed grid model; s2, inputting the preprocessed grid model and secret information into a secret information embedding algorithm based on vertex division and smooth ordering for embedding processing to obtain a load-dense grid model; s3, the receiver receives the dense grid model, and performs data type conversion on the vertex coordinates of the dense grid model to obtain a preprocessed dense grid model; S4, inputting the preprocessed dense grid model into a secret information extraction algorithm based on vertex division and smooth sorting for extraction processing, and obtaining an original grid model and secret information. Further, the process of embedding the secret information embedding algorithm based on vertex division and smooth ordering is as follows: S21, dividing all vertexes in the preprocessed grid model to obtain a candidate point set and a reference point set; S22, acquiring a central vertex according to the candidate point set, acquiring a ring of adjacent reference