Unified Scrambling And Reversible Data Hiding For Three-Dimensional Mesh Models

The widespread use of 3D models and the rapid development of cloud storage have made 3D models highly marketable,but they have also brought security issues.Attackers can easily copy 3D models from cloud storage,which can lead to copyright infringement of 3D models.To ensure security,3D models are usually encrypted before uploading to the cloud,this requires a reversible data hiding RDH technique.The original encryption domain and reversible data hiding RDH were to solve different multimedia security problems.The purpose of encryption was to protect the content in the image signal.The purpose of reversible data hiding RDH is to imperceptibly and reversibly embed secret information into multimedia signals,but existing reversible data hiding RDH cannot be directly applied to 3D mesh models.Because the mesh model is characterized by vertices and faces rather than colors.Although there are some techniques available for 3D models on reversible data hiding RDH,it cannot be directly applied to encrypted 3D mesh models.For encrypted 3D mesh models,the correlation between adjacent vertex coordinates is corrupted.Therefore,one cannot predict the adjacent coordinates of the encrypted 3D mesh.This problem forced researchers to develop reversible data hiding RDH algorithms for encrypted 3D meshes.But in the existing algorithms,the scrambling and data hiding are separated,not jointly designed.Although data embedding does not affect the restoration of the original mesh,it does affect the embedding capacity during the data hiding phase.After the data is encrypted,the correlation is destroyed,thereby reducing the available redundancy.Therefore,the hidden capacity is also reduced.This cannot meet the requirements of high scrambling rate and high embedding rate data hiding at the same time.In view of the above problems,we propose an adaptive unified scrambling and reversible data hiding algorithm(3D-AUES)for the 3D mesh model.This technology can simultaneously embed external data and scramble 3D meshes.First,we adopt a vertex coordinate prediction method called cross prediction to accurately predict the other half of the vertices.The predicted vertices are used to embed external data.Then we embed the external data by bit replacement.This step further improves the embedding rate.To improve security,we propose an adaptive threshold method to select vertices to embed.To ensure lossless scrambling,the threshold and prediction error are embedded into the vertex coordinates as auxiliary information and secret information.By adopting an adaptive threshold and multilayer embedding,scalable scrambling quality can be achieved.At the decoding end,using the lossless embedded auxiliary information,external data can be successfully extracted,and the mesh can be restored to a predetermined level of distortion,which can achieve recovery from lossless to partial recovery.The experimental results show that the unified data scrambling and reversible data hiding technology(3D-AUES)uses a bit replacement scheme to embed external data,which can achieve a higher embedding rate.Adaptive control threshold parameters can achieve scalable scrambling quality and scalable recovery quality.