Superior SOM Neural Network based Minute Significant Watermark Generator and Detector System
N. Chenthalir Indra, E. Ramaraj · International Journal of Computer Applications · 2011
This paper suggests the Superior SOM (SSOM) based Minute Significant Watermark Generator & Detector (MSWG&D) system.RGB features of the host image are trained in different SSOM networks.Subsequent to SSOM training process, microscopic significant values are synthesized from host image and self-possessed as watermark values.Then these values are embedded into the high frequency sub band of Discrete Wavelet Transform (DWT).The Quality of invisible watermarking is proved by evaluating PSNR & Jaccard Similarity Ratio values between original and watermarked image.MSWG&D system is robust to JPEG compression and noise attacks.The experimental results prove that the strength of proposed watermarking system is 'one more landmark' in the watermarking techniques.