Image Security Enhancement By Semi Fragile Watermarking Appended With Unsharp Masking

Abstract

In present era of secure communication, various authentication techniques are used to protect or to resolve the copyright integrity of audio, video and multimedia data files. Digital watermarking is one of the most efficient ways to protect the digital properties of multimedia files in which the original data files to be transferred are embedded with some specific information which is hard to detect and remove. The watermark should be embedded into the host image in a way that it should not degrade the quality of the original watermarked image. Digital image watermarking finds applications in varied areas like defence, medical science, intellectual property right and entertainment. Watermarked digital image is transmitted or stored, where it can suffer from distinct attacks. There are many promising modifications, such as cropping an image, lossy compression of the information or adding noises. So protections against such attacks should be required. In the proposed semi fragile watermarking technique LBP (Local Binary Pattern) is used for embedding watermark. Besides being able to detect all the malicious changes that affect the quality of watermarked image, it is tolerant to image compression which is essential to transmit images over the network. In this research work, a framework for enhancement of watermarked image quality using LBP (Local Binary Pattern) watermarking and image enhancement is proposed by unsharp masking. Overall the image quality is enhanced as the PSNR calculated using LBP alone is 45.0398 and using proposed algorithm the value has improved to 66.2204. Moreover, the proposed method is robust against some commonly used image processing operations such compression, contrast, noise, tampering and crop attack providing PSNR of 56.6190, 50.0969, 51.0631, 41.0486 and 43.6339 respectively with MSE of 0.0560, 0.3336, 0.2245, 0.7674 and 0.6781also when compared with the traditional method there is improvement of 22.7620% in PSNR when subjected to additive noise attack, 23.8817% improvement in PSNR when subjected to contrast adjustment, compression with 18.651% improvement in PSNR under compression attack. This technique provides good results against cropping and tampering attacks along with improvement in PSNR of 32.223% and 32.9692% respectively with least computational cost.