Effects of Image Compression on Image Age Approximation

In-field sensor defects are at the core of temporal image forensics, as a temporal order among pieces of evidence can be established by knowing their onset time. A characteristic of these single pixel defects is that they appear as point-like image noise. Since sensor defects exhibit noise-like properties, they are vulnerable to image compression. For this reason, it is important to evaluate the effect of image compression on the defect based image age approximation techniques available. In this paper, we assess the robustness of these age approximation methods with respect to four different compression techniques (i.e., 'JPEG', 'JPEG 2000', 'JPEG XR' and 'Better Portable Graphics') and different compression strengths. Since the approximation techniques considered require the defect locations to be known in advance, we assess the effect of image compression on the defect detection methods proposed in the context of image age approximation.