An Adaptive High-Capacity Steganographic Algorithm for 3D Models

Abstract

In this paper, a new simple, adaptive, high-capacity steganographic algorithm for 3D models using contour theory in spatial domain is proposed. Every vertex of a 3D model can adaptively embed variable 3σ ( 1 σ ≥ ) bits using contour space subdivision and multi quantization index modulation (MQIM) with low distortion. We first use vertex weight estimation which is used to estimate the degree of smoothness or roughness properties of 3D model with respect to human visual system and adaptive embedding estimation to provide adaptive and high capacity, then contour space subdivision is used to project the vertex to contour plane, finally at least four bits can be embedded by multi QIM in a vertex. The experiment results show that the proposed technique is simple, adaptive and secure, has high capacity and low distortion, and is robust against affine transformations and vertex reordering attacks, so it is suitable for steganography of arbitrary 3D mesh models.