Stay-Positive: A Case for Ignoring Real Image Features in Fake Image Detection

Anirudh Sundara Rajan, Yong Jae Lee
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:58008-58028, 2025.

Abstract

Detecting AI-generated images is a challenging yet essential task. A primary difficulty arises from the detector’s tendency to rely on spurious patterns, such as compression artifacts, which can influence its decisions. These issues often stem from specific patterns that the detector associates with the real data distribution, making it difficult to isolate the actual generative traces. We argue that an image should be classified as fake if and only if it contains artifacts introduced by the generative model. Based on this premise, we propose Stay-Positive, an algorithm designed to constrain the detector’s focus to generative artifacts while disregarding those associated with real data. Experimental results demonstrate that detectors trained with Stay-Positive exhibit reduced susceptibility to spurious correlations, leading to improved generalization and robustness to post-processing. Additionally, unlike detectors that associate artifacts with real images, those that focus purely on fake artifacts are better at detecting inpainted real images.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-sundara-rajan25a, title = {Stay-Positive: A Case for Ignoring Real Image Features in Fake Image Detection}, author = {Sundara Rajan, Anirudh and Lee, Yong Jae}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {58008--58028}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/sundara-rajan25a/sundara-rajan25a.pdf}, url = {https://proceedings.mlr.press/v267/sundara-rajan25a.html}, abstract = {Detecting AI-generated images is a challenging yet essential task. A primary difficulty arises from the detector’s tendency to rely on spurious patterns, such as compression artifacts, which can influence its decisions. These issues often stem from specific patterns that the detector associates with the real data distribution, making it difficult to isolate the actual generative traces. We argue that an image should be classified as fake if and only if it contains artifacts introduced by the generative model. Based on this premise, we propose Stay-Positive, an algorithm designed to constrain the detector’s focus to generative artifacts while disregarding those associated with real data. Experimental results demonstrate that detectors trained with Stay-Positive exhibit reduced susceptibility to spurious correlations, leading to improved generalization and robustness to post-processing. Additionally, unlike detectors that associate artifacts with real images, those that focus purely on fake artifacts are better at detecting inpainted real images.} }
Endnote
%0 Conference Paper %T Stay-Positive: A Case for Ignoring Real Image Features in Fake Image Detection %A Anirudh Sundara Rajan %A Yong Jae Lee %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-sundara-rajan25a %I PMLR %P 58008--58028 %U https://proceedings.mlr.press/v267/sundara-rajan25a.html %V 267 %X Detecting AI-generated images is a challenging yet essential task. A primary difficulty arises from the detector’s tendency to rely on spurious patterns, such as compression artifacts, which can influence its decisions. These issues often stem from specific patterns that the detector associates with the real data distribution, making it difficult to isolate the actual generative traces. We argue that an image should be classified as fake if and only if it contains artifacts introduced by the generative model. Based on this premise, we propose Stay-Positive, an algorithm designed to constrain the detector’s focus to generative artifacts while disregarding those associated with real data. Experimental results demonstrate that detectors trained with Stay-Positive exhibit reduced susceptibility to spurious correlations, leading to improved generalization and robustness to post-processing. Additionally, unlike detectors that associate artifacts with real images, those that focus purely on fake artifacts are better at detecting inpainted real images.
APA
Sundara Rajan, A. & Lee, Y.J.. (2025). Stay-Positive: A Case for Ignoring Real Image Features in Fake Image Detection. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:58008-58028 Available from https://proceedings.mlr.press/v267/sundara-rajan25a.html.

Related Material