Hypothesis-based vertex shift method for embedding secret logos in the geometric features of 3D objects

Authors: ARTHY R, MALA K, SURESH BABU R

Abstract: A recent challenge in information technology is to protect secret data and preserve the ownership of a product. There are many duplicate products being released on a daily basis. Owners have a high risk in proving their products. Watermarking is a technique used to preserve ownership by hiding the owner's information in their products. The proposed hypothesis-based vertex shifting algorithm embeds 2D secret logos in 3D cover objects. The 3D objects are represented using vertices and facets. 3D watermarking faces various challenges and one among them is capacity. In this work, capacity is addressed by using a hypothesis-based vertex shift method that enables the embedding process for all the coordinates of the vertex. The method works by partitioning the vertex based on a shift factor called svalue. The svalue is chosen based on the visual quality of the watermarked object. The metrics used for testing are bit error rate for the recovered watermark, peak-to-signal noise ratio, and vertex signal-to-noise ratio (VSNR) of the watermarked 3D image. The proposed algorithm shows that a maximum of 3 bits can be embedded in a vertex when compared with the existing algorithms. The VSNR value of the proposed algorithm is high (125.87) compared to the existing algorithms. This shows that the algorithm withstands visual quality inspection. Hence, it is a robust watermarking algorithm for embedding secret logos into 3D objects with better visual quality and higher resilience against translation and uniform scaling attacks.

Keywords: 3D objects, watermarking, peak signal-to-noise ratio, bit error rate, vertex signal-to-noise ratio, attack, translation attack, uniform scaling attack

Full Text: PDF