This thesis presents a reversible data hiding method which is affine transformation invariant for 3D mesh models. The histogram shift scheme is applied to data embedding by modifying the normalized distances between the model center and vertices. When the 3D model is rotated, uniform scaled, or translated, the embedded data can be correctly extracted and the original 3D model can be reconstructed. For blind extraction of hidden data, there is only a small extra payload that should be recorded. Although the 3D model is slight changed by modifying vertex coordinates, it cannot be detected by human eyes. The experimental results show that the proposed method is robust against the rotation, uniform scaling, and translation attacks.
