Title: Sub-band discrete cosine transform-based greyscale image watermarking using general regression neural network

Authors: Rajesh Mehta; Navin Rajpal; Virendra P. Vishwakarma

Addresses: University School of Information and Communication Technology, Guru Gobind Singh Indraprastha University, Sector 16-C, Dwarka, New Delhi, India ' University School of Information and Communication Technology, Guru Gobind Singh Indraprastha University, Sector 16-C, Dwarka, New Delhi, India ' University School of Information and Communication Technology, Guru Gobind Singh Indraprastha University, Sector 16-C, Dwarka, New Delhi, India

Abstract: In this paper, a new grey scale image watermarking scheme based on sub-band discrete Cosine transform (SB-DCT) using general regression neural network (GRNN) is proposed. The image features are extracted by applying the SB-DCT to each non-overlapping block of the image. These features are used to form the dataset, which act as input to GRNN. The output obtained by GRNN is used to embed the binary watermark logo in the selected low variance blocks of the image. Owing to the good function approximation and high generalisation property of GRNN, we are able to recover the watermark after performing several image processing operations. Through the extensive experimental results, high peak signal-to-noise ratio (PSNR) value of watermarked image and high bit correct ratio (BCR), normalised correlation (NC) value of the extracted watermark proves the imperceptibility and robustness of the proposed scheme compared to the state-of-art techniques.

Keywords: bit correct ratio; BCR; discrete cosine transform; DCT; general regression neural networks; GRNNs; normalised correlation; sub-band decomposition; greyscale image watermarking; image features; feature extraction; peak SNR; signal-to-noise ratio; PSNR.

DOI: 10.1504/IJSISE.2015.072927

International Journal of Signal and Imaging Systems Engineering, 2015 Vol.8 No.6, pp.380 - 389

Received: 14 Feb 2014
Accepted: 07 Jul 2014
Published online: 08 Nov 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article