Reversible Data Hiding Based on Octree Partitioning and Arithmetic Coding in Encrypted Three-Dimensional Mesh Models

Xiaochun Wu, Jiawang Zhang, Qi Yao, Wanli Lyu

Abstract: Reversible data hiding in encrypted domain technology embeds additional information in encrypted carriers, ensuring the privacy and security of both the carriers and the additional information during transmission. However, existing researches in this area has mainly focused on images, and there is an urgent need to improve data embedding rate for three-dimensional models while ensuring reversibility. In this paper, we put forward a method built on octree space partitioning and arithmetic coding for 3D mesh models. Initially, the octree space partitioning strategy is employed to divide the vertices into disjoint subblocks. Within each subblock, the vertices are categorized into an embedding set and a reference set. Subsequently, additional data is embedded into the embedding set. The encrypted model is then losslessly recovered using the reference set. To reduce the storage cost of auxiliary information, a uniform minimum prediction length is applied across three coordinate axes, and arithmetic coding is used for compression. The results of the experiment demonstrate that this method maintains high embedding ability, guarantees lossless model recovery, and effectively reduces the computational complexity.

External IDs:dblp:conf/icic/WuZYL24