Chrominance based Splicing Image Forgery Localization with Relevance Vector Machine

Due to the modern technology, image forgery is very much prevalent nowadays. Image forgery may result to misleading facts if used without proper verification. Thus, image forgery detection plays a very vital role. The image forgery detection techniques confirm the credibility of the digital images with no prior information about the original image. Since, the luminance component of the image is perceived by humans, tampering may result in some unnatural clues in the chrominance component. In this paper, the goal is to detect and localize the digital images tampered using image splicing techniques by analyzing the chrominance component. The inconsistencies in the noise level in a image due to tampering is evaluated and used as the feature to classify using Relevance Vector Machine (RVM). The proposed method obtains an accuracy of 98.75% for Cb channel while 99.02% accuracy is obtained for Cr channel which is the highest accuracy among all existing methods.