Image aurhentication using forensic techniques

Abstract

The impact of technology depends on the users. The worldwide web and photo editing software are designed and devices for benefiting the society. However, powerful photo-editing software are used for creating forgery in images and worldwide web is used for spreading the information without checking the authenticity of information. In the age of social media and wide spread and accessibility of worldwide web, the spread of false information in form of images becomes very easy. The authenticity determination can avoid spread of false information and playing important role in decision making, where images are used as evidence for e.g. in a court of law, photographic composites of human faces, political events, court evidence, financial contracts, scientific discoveries, news reports and other fields leading to social severe consequences. newlineCopy move forgery, which involves the copying any object and pasting in same image with preprocessing and post processing operations to create forgery, is the most popular and difficult in detection due to simile properties of source object. Based on the comprehensive literature the copy move forgery detection (CMFD) are classified in brute force, autocorrelation, block based and key point based algorithms. The brute force and autocorrelation based CMFD techniques are abundant due to exhaustive searching and size of copied region. The block based CMFD algorithms have high computational complexity, non-invariance to all attack types, block size selection problem leading to high true negatives and false positives. The key point algorithms performance is very poor in smooth regions and have high false positives. Block based CMFD algorithms perform better in smooth regions and key point based CMFD has better computation time performance, invariance to various attack types leading to formulation of hybrid techniques involving both block based and key point based CMFD algorithms. The research gap of performance improvement in hybrid techniques in extreme smooth and extreme textured images