Papers › Learning JPEG Compression Artifacts for Image Manipulation Detection and Localization

Learning JPEG Compression Artifacts for Image Manipulation Detection and Localization

30 Aug 2021arXiv:2108.12947archive 2025-07-28

Myung-Joon Kwon, Seung-Hun Nam, In-Jae Yu, Heung-Kyu Lee, Changick Kim

Detecting and localizing image manipulation are necessary to counter malicious use of image editing techniques. Accordingly, it is essential to distinguish between authentic and tampered regions by analyzing intrinsic statistics in an image. We focus on JPEG compression artifacts left during image acquisition and editing. We propose a convolutional neural network (CNN) that uses discrete cosine transform (DCT) coefficients, where compression artifacts remain, to localize image manipulation. Standard CNNs cannot learn the distribution of DCT coefficients because the convolution throws away the spatial coordinates, which are essential for DCT coefficients. We illustrate how to design and train a neural network that can learn the distribution of DCT coefficients. Furthermore, we introduce Compression Artifact Tracing Network (CAT-Net) that jointly uses image acquisition artifacts and compression artifacts. It significantly outperforms traditional and deep neural network-based methods in detecting and localizing tampered regions.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

mjkwon2021/cat-net officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Image ManipulationImage Manipulation DetectionImage Manipulation Localization

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Manipulation Detection COVERAGE CAT-Net v2 AUC .680 #4 of 8 Archive leaderboard report
Image Manipulation Detection COVERAGE CAT-Net v2 Balanced Accuracy .635 #4 of 8 Archive leaderboard report
Image Manipulation Detection Casia V1+ CAT-Net v2 AUC .942 #3 of 9 Archive leaderboard report
Image Manipulation Detection Casia V1+ CAT-Net v2 Balanced Accuracy .838 #3 of 9 Archive leaderboard report
Image Manipulation Detection CocoGlide CAT-Net v2 AUC .667 #4 of 8 Archive leaderboard report
Image Manipulation Detection CocoGlide CAT-Net v2 Balanced Accuracy .580 #4 of 8 Archive leaderboard report
Image Manipulation Detection Columbia CAT-Net v2 AUC .977 #4 of 8 Archive leaderboard report
Image Manipulation Detection Columbia CAT-Net v2 Balanced Accuracy .803 #4 of 8 Archive leaderboard report
Image Manipulation Detection DSO-1 CAT-Net v2 AUC .747 #4 of 9 Archive leaderboard report
Image Manipulation Detection DSO-1 CAT-Net v2 Balanced Accuracy .525 #4 of 9 Archive leaderboard report
Image Manipulation Localization CASIAv1(Protoclo-CAT) CAT-Net Pixel Binary F1 0.808 #3 of 8 Archive leaderboard report
Image Manipulation Localization COVERAGE CAT-Net v2 Average Pixel F1(Fixed threshold) .381 #9 of 11 Archive leaderboard report
Image Manipulation Localization COVERAGE(Protocol-CAT) CAT-Net Pixel Binary F1 0.427 #5 of 8 Archive leaderboard report
Image Manipulation Localization Casia V1+ CAT-Net v2 Average Pixel F1(Fixed threshold) .752 #6 of 11 Archive leaderboard report
Image Manipulation Localization CocoGlide CAT-Net v2 Average Pixel F1(Fixed threshold) .434 #10 of 11 Archive leaderboard report
Image Manipulation Localization Columbia CAT-Net v2 Average Pixel F1(Fixed threshold) .859 #6 of 11 Archive leaderboard report
Image Manipulation Localization Columbia(Protocol-CAT) CAT-Net Pixel Binary F1 0.915 #3 of 8 Archive leaderboard report
Image Manipulation Localization DSO-1 CAT-Net v2 Average Pixel F1(Fixed threshold) .584 #7 of 11 Archive leaderboard report
Image Manipulation Localization NIST16(Protocol-CAT) CAT-Net Pixel Binary F1 0.252 #6 of 8 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

ConvolutionDiscrete Cosine Transform

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections