Performance analysis of meta-heuristics on dual watermarking of color images based on SWT and SVD

This paper proposes a robust and high embedding capacity watermarking technique where two watermarks are embedded in a color image. A gray-scale image is embedded as one of the watermarks to identify the ownership of the host image using Stationary Wavelet Transform (SWT) and Singular Value Decomposition (SVD). A binary watermark with the help of the least significant bit method is embedded to locate the tampers. This technique is optimized to have better robustness against various attacks by employing meta-heuristics. The performance of the Genetic Algorithm (GA), Invasive Weed Optimization (IWO) and Teaching Learning Based Optimization (TLBO) is analyzed on the proposed watermarking technique. A novel fitness function is proposed to optimize color image watermarking. This technique is applied to standard images and medical images. The performance of the meta-heuristics on the proposed method is compared with respect to the imperceptibility, robustness, the number of pixels changing rate (NPCR) and computation complexity.