Abstract:
Rapid advancement in digital image processing tools and software’s has made it extremely
simple to manipulate the digital images without leaving any footprints. It becomes a hot issue about the
security and threat to society with increasing growth of social media. JPEG compression format has been
widely used in most of the digital cameras. The investigation of JPEG compression footprints can play
an important role in image tampering detection. In this paper, a novel method is proposed to detect the
JPEG compression. The proposed forensic scheme comprises of two steps i.e. selection of target difference
image and generation of second-order statistical features by evaluating the Markov Transition Probability
Matrices (MTPMs) for both intra and inter-block DCT domain. Finally, the resultant feature is used to train
the SVM classifier for classification purposes. The experiment results on UCID and BOSSBase datasets
show that the proposed forensic technique based on MTPM is capable of detecting the JPEG compression
traces even in the presence of anti-forensic attacks.
