Logo embedding in grayscale images using turbo product codes

Abstract

In this paper, we investigate the robustness achieved by hard-input hard-output (HIHO) turbo product code (TPC) in a grayscale image watermarking. The TPC scheme is based on the concatenation of a BCH product code. The general concept of the TPC is to construct two or more simple codes in order to achieve remarkable error performance. Unlike many proposed algorithms in the literature, HIHO-TPC require a manageable complexity encoding and decoding algorithms. Firstly, the data in the logotype (logo) is encoded using TPC and then inserted in the image. Different grayscale images are investigated and evaluated in terms of robustness to different photometric attacks by using Stirmark software package. Experimental results show that the proposed blind watermarking algorithm has a strong ability of extracting the embedded logo in different attacks such as JPEG compression, cropping, and 3x3 median filtering.