A study of the robustness of PRNU-based camera identification

We investigate the robustness of PRNU-based camera identification in cases where the test images have been passed through common image processing operations. A specific question we aim to answer is how camera identification can be circumvented by an nontechnical user, applying only standard and/or freely available software. We study denoising, recompression, out-of-camera demosaicing. We measure the beneficial effect of JPEG block artifact removal on camera identification accuracy. We examine the extent to which JPEG block artifact removal helps circumvention by improving the performance of a denoiser that is used to remove PRNU features. We evaluate the effectiveness of the StirMark watermark remover as a tool for circumventing camera identification, and we examine the similarities and differences between watermark circumvention and camera identification circumvention.