Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
405842 | Neurocomputing | 2016 | 6 Pages |
Image anti-forensics, which aims to remove or forge traces upon which image forensics is based, has made rapid progress recently. To rebuild the credibility of forensics, many countermeasures have been proposed for detecting different anti-forensics. However, most existing countermeasures just target only one type of anti-forensics and are difficult to extend to counter other anti-forensics. In this paper, a multi-purpose countermeasure using autoregressive (AR) model is proposed for detecting various anti-forensics. Experimental results demonstrate that the proposed countermeasure achieves satisfactory performance in detecting all of the five well-known anti-forensic methods discussed in this paper. Even compared to the state-of-art specific counter-measures, our proposed countermeasure achieves similar or better performance.