A Cluster－Based Multiple Watermarking Algorithm for Compound 3D Models

Multimedia watermarking techniques are used to embed extra information into multimedia contents (e.g. images, audio signals, and 3D models). Frequently, embedding of multiple watermarks in a single multimedia content is required. When a multimedia content is a common work created together by several authors, it might be necessary to embed each co-author’s watermark into the common multimedia content. For this purpose, a watermarking technique capable of embedding multiple watermarks for compound 3D model based on clustering is proposed in this paper. In the proposed technique, the original compound 3D model is clustered into several disconnected simple 3D models by means of a vertices connectivity filter. These watermarks are then simultaneously embedded into individual clusters. Imperceptibility of the watermarks was confirmed through the computation of SNR of the watermarks embedded 3D model. Robustness against several kinds of geometric attacks was evaluated by NC, SCC, and BER value of extracted watermarks. The experiment results show that the proposed algorithm can efficiently embedseveral watermarks simultaneously into compound 3D models, and the embedded watermark is robust against to rotation, translation, and uniformly scaling attacks.