PiD: Generalized AI-Generated Images Detection with Pixelwise Decomposition Residuals

Xinghe Fu, Zhiyuan Yan, Zheng Yang, Taiping Yao, Yandan Zhao, Shouhong Ding, Xi Li
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:17894-17908, 2025.

Abstract

Fake images, created by recently advanced generative models, have become increasingly indistinguishable from real ones, making their detection crucial, urgent, and challenging. This paper introduces PiD (Pixelwise Decomposition Residuals), a novel detection method that focuses on residual signals within images. Generative models are designed to optimize high-level semantic content (principal components), often overlooking low-level signals (residual components). PiD leverages this observation by disentangling residual components from images, encouraging the model to uncover more underlying and general forgery clues independent of semantic content. Compared to prior approaches that rely on reconstruction techniques or high-frequency information, PiD is computationally efficient and does not rely on any generative models for reconstruction. Specifically, PiD operates at the pixel level, mapping the pixel vector to another color space (e.g., YUV) and then quantizing the vector. The pixel vector is mapped back to the RGB space and the quantization loss is taken as the residual for AIGC detection. Our experiment results are striking and highly surprising: PiD achieves 98% accuracy on the widely used GenImage benchmark, highlighting the effectiveness and generalization performance.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-fu25i, title = {{P}i{D}: Generalized {AI}-Generated Images Detection with Pixelwise Decomposition Residuals}, author = {Fu, Xinghe and Yan, Zhiyuan and Yang, Zheng and Yao, Taiping and Zhao, Yandan and Ding, Shouhong and Li, Xi}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {17894--17908}, 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/fu25i/fu25i.pdf}, url = {https://proceedings.mlr.press/v267/fu25i.html}, abstract = {Fake images, created by recently advanced generative models, have become increasingly indistinguishable from real ones, making their detection crucial, urgent, and challenging. This paper introduces PiD (Pixelwise Decomposition Residuals), a novel detection method that focuses on residual signals within images. Generative models are designed to optimize high-level semantic content (principal components), often overlooking low-level signals (residual components). PiD leverages this observation by disentangling residual components from images, encouraging the model to uncover more underlying and general forgery clues independent of semantic content. Compared to prior approaches that rely on reconstruction techniques or high-frequency information, PiD is computationally efficient and does not rely on any generative models for reconstruction. Specifically, PiD operates at the pixel level, mapping the pixel vector to another color space (e.g., YUV) and then quantizing the vector. The pixel vector is mapped back to the RGB space and the quantization loss is taken as the residual for AIGC detection. Our experiment results are striking and highly surprising: PiD achieves 98% accuracy on the widely used GenImage benchmark, highlighting the effectiveness and generalization performance.} }
Endnote
%0 Conference Paper %T PiD: Generalized AI-Generated Images Detection with Pixelwise Decomposition Residuals %A Xinghe Fu %A Zhiyuan Yan %A Zheng Yang %A Taiping Yao %A Yandan Zhao %A Shouhong Ding %A Xi Li %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-fu25i %I PMLR %P 17894--17908 %U https://proceedings.mlr.press/v267/fu25i.html %V 267 %X Fake images, created by recently advanced generative models, have become increasingly indistinguishable from real ones, making their detection crucial, urgent, and challenging. This paper introduces PiD (Pixelwise Decomposition Residuals), a novel detection method that focuses on residual signals within images. Generative models are designed to optimize high-level semantic content (principal components), often overlooking low-level signals (residual components). PiD leverages this observation by disentangling residual components from images, encouraging the model to uncover more underlying and general forgery clues independent of semantic content. Compared to prior approaches that rely on reconstruction techniques or high-frequency information, PiD is computationally efficient and does not rely on any generative models for reconstruction. Specifically, PiD operates at the pixel level, mapping the pixel vector to another color space (e.g., YUV) and then quantizing the vector. The pixel vector is mapped back to the RGB space and the quantization loss is taken as the residual for AIGC detection. Our experiment results are striking and highly surprising: PiD achieves 98% accuracy on the widely used GenImage benchmark, highlighting the effectiveness and generalization performance.
APA
Fu, X., Yan, Z., Yang, Z., Yao, T., Zhao, Y., Ding, S. & Li, X.. (2025). PiD: Generalized AI-Generated Images Detection with Pixelwise Decomposition Residuals. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:17894-17908 Available from https://proceedings.mlr.press/v267/fu25i.html.

Related Material