Performance Comparison of Product Codes and Cubic Product Codes using FRBS for Robust Watermarking

Abstract

In this paper, a novel technique for robust digital image watermarking is proposed using simple product codes (PC) and cubic product codes (CPC) respectively. Product codes are two dimentional matrix codes while CPC are three dimensional product codes where the constituent codes in each dimension are linear block codes. The structure of PC and CPC makes them suitable for the proposed scheme. The embedded watermark is vulnerable to various attacks on the image like compression, noise and geometric attacks namely translation rotation and scaling (TRS) etc. This could limit the performance of digital watermarking schemes. Our proposal is to encode the watermark with PC/CPC prior to embedding it into the image. This could easily be done because our watermark is also a three dimensional data (image/logo etc), so each dimension can easily be encoded with corresponding codes in PC/CPC. The modified iterative decoding algorithm (MIDA) is employed to decode the PC/CPC. Moreover, a Fuzzy Rule Based System (FRBS) is used to find the suitable regions in the host image where watermark can be embedded such that the imperceptibilityof the host image should not be affected. The FRBS makes used of Human Visual System (HVS) parameters to obtain those regions. The scheme is tested against various attacks and compared with the well-known schemes in the literature. Moreover, role of PC is also compared with CPCs and results are demonstrated.