Current secret sharing techniques are mostly proposed on digital images. However, extending the secret sharing schemes to 3D models is necessary because of rapid development of 3D multimedia techniques. Note that the basic composition elements and the range of values between 3D models and digital images are very different. Therefore, we will propose a novel 3D secret sharing algorithm considering the characteristics of 3D models. This project will be carried out in two years. In the first year of this project, we will propose a novel 3D secret sharing scheme based on the space subdivision technique. Each vertex within the secret point geometry is firstly encoded into a series of integer values. The share values are then derived from the share equation whose coefficients are replaced by above encoded values. Finally, each participate will separately have a 3D stego model with the share values embedded. Our proposed technique has the following advantages, including higher privacy for the share values, higher robustness and lower distortion for the stego model and higher efficiency for our proposed algorithm. Sharing a polygonal model is difficult, however, because of the different data types for its vertex coordinates and polygon indices. The degree of difficult can be more increased when the cover model used to embed the share values is a polygonal one. Therefore, in the second year of this project, we will continue the achievements of the first year to extend the previous algorithm to support the polygonal model as the secret model. A novel progressive model reconstruction technique is proposed in the sampling domain by integrating our previous plan in the last year. We believe that this algorithm will do a great contribution to the area of 3D secret sharing.
