Highlighting image tampering by feature extraction based on image quality deterioration

Because of advancement in information technology and image processing software, Image Tampering and its forensics is gaining research interest. This paper aims at detecting image manipulations by exploiting the noise discrepancies which arise due to image quality deterioration. RGB, YCbCr and L*a*b color model are used to check the quality deterioration. Characteristics based features are extracted and then utilized for differentiating original image from manipulated one. Experiments proved that features utilizing various color models together to highlight noise discrepancies can classify the image better as compared to techniques considering individual color model.