Statistical Feature based Blind Classifier for JPEG Image Splice Detection

Digital imaging, image forgery and its forensics have become an established field of research now days. Digital imaging is used to enhance and restore images to make them more meaningful while image forgery is done to produce fake facts by tampering images. Digital forensics is then required to examine the questioned images and classify them as authentic or tampered. This paper aims to design and implement a blind classifier to classify original and spliced Joint Photographic Experts Group (JPEG) images. Classifier is based on statistical features obtained by exploiting image compression artifacts which are extracted as Blocking Artifact Characteristics Matrix. The experimental results have shown that the proposed classifier outperforms the existing one. It gives improved performance in terms of accuracy and area under curve while classifying images. It supports .bmp and .tiff file formats and is fairly robust to noise