Design And Research Of 3D Point Cloud Model Information Hiding Algorithm Based On Plaintext Domain And Encrypted Domain

As a new digital medium,3D models have increasingly wide application fields;and information hiding,as a digital medium-based information security technology,can effectively not protect the information security of 3D models,and information hiding algorithms that choose 3D models as carriers are emerging research fields.For this reason,the thesis selects3 D point cloud model as information hiding carrier,and proposes plaintext domain information hiding algorithm to provide more robust protection for secret information for the copyright protection scenario of multimedia data;and proposes encrypted domain reversible information hiding algorithm to provide double guarantee for the security of model itself content and secret information for the scenario that needs model protection and lossless recovery of original model.The main research work and results of the paper are as follows:(1)A plaintext domain information hiding algorithm based on multiscale transformation and composite operator is proposed.Firstly,the point cloud model is rasterized,the corner points of the 3D point cloud model are extracted using the improved 3D Harris algorithm,and the robust feature points are optimally screened as the embedding region of the secret information by combining with the SURF operator;finally,the multi-scale transformation is performed on the feature region,and the secret information is embedded into the low-frequency coefficient matrix by using the quantization-based method to achieve the hiding of the secret information.The experimental results show that the algorithm can reach a Corr value of 0.729 in the face of the compound attack of 10% simplified 0.5% noise and 10% shear,which has good invisibility and strong robustness to affine transformation attack compared with many algorithms of the same type.(2)Proposed reversible information hiding algorithm for cryptographic domain point cloud models based on prediction error expansion.First,before the point cloud model is stored to the cloud,it is encrypted using a new chaotic system to prevent extraneous people from accessing the content;second,the vertices of the model are classified using a greedy algorithm to calculate the prediction error of the vertices and extend their mode length to embed the secret information;finally,at the receiving end,the secret information is extracted and the original point cloud model is recovered by comparing the mode length of the prediction error.The experimental results show that the embedding performance of the algorithm is significantly improved,and the average embedding rate is improved by 0.284 bpv and 0.298 bpv,respectively,compared with the comparison algorithm,and the model is recovered losslessly to ensure the reversibility of the algorithm.In summary,the proposed information hiding algorithms for different scenarios outperform a variety of existing mainstream algorithms and can effectively meet the needs of different application scenarios with good application prospects.