Comparative Performance Analysis of Color Image watermarking Scheme Using Hessenberg Decomposition and Schur Decomposition

摘要

This blind color image watermarking technique based on Hessenberg Decomposition and Schur Decomposition examines the performance in terms of invisibility and robustness. For the security purpose, the Fibonacci-Lucas Transform is used to scramble the watermark image. This technique is tested with several color images from the image database CVG-UGR as the host image and QR code image is used as a binary watermark image. In this method the Y component of the YCbCr color space of the color host image is used to embed the binary QR code as the watermark image. In term of average PSNR value i.e. 64.1177 dB of all the watermarked images, better Perceptual quality achieved using Hessenberg Decomposition as compared to the average PSNR value 55.5823 dB using Schur Decomposition. Moreover, all the recovered watermark image i.e. QR codes have been detected too. The analysis shows good robustness against several attacks. Hessenberg Decomposition satisfies the good robustness against several noise like salt and pepper noise, resize, cropping, gaussian noise, resize and JPEG compression as compared to Schur Decomposition.