Abstract:Reversible data hiding in the encrypted domain (RDHED) enables the embedding of secret data without revealing the privacy of the carrier. However, in view of the generally low embedding capacity of existing RDHED methods for 3D mesh models, this study proposes a reversible data hiding method for encrypted 3D mesh models based on octree partitioning and a vertex segmentation strategy. Firstly, the model is adaptively divided into non-overlapping blocks, with intra-block spatial correlation preserved. Then, a partitioning strategy based on vortex entropy is designed to accurately select reference points for prediction accuracy enhancement. Finally, an adaptive most significant bit (MSB) prediction method is employed to maximize the embeddable space of each vortex and thus notably improve the embedding capacity. Experimental results demonstrate that the proposed method improves the embedding capacity of 3D mesh models while ensuring data reversibility and separability, thus offering an effective solution for reversible data hiding in 3D models.