Diffusion-based Semantic Outlier Generation via Nuisance Awareness for Out-of-Distribution Detection
Suhee Yoon, Sanghyu Yoon, Ye Seul Sim, Sungik Choi, Kyungeun Lee, Hye-Seung Cho, Hankook Lee, Woohyung Lim
Abstract
Out-of-distribution (OOD) detection, determining whether a given sample is part of the in-distribution (ID) or not, has been newly explored by a generative model-based outlier synthesizing approach, especially with diffusion models. Nonetheless, existing diffusion models often produce outliers that are considerably distant from the ID in pixel-space, showing limited efficacy for capturing subtle distinctions between ID and OOD. To address these issues, we propose a novel framework, Semantic Outlier generation via Nuisance Awareness (SONA), which directly utilizes informative pixel-space ID images in diffusion models. Thereby, the generated outliers achieve two crucial properties: (i) they closely resemble the ID mainly in nuisances, while (ii) represent discriminative semantic information. To facilitate the separate effect on semantics and nuisances, we introduce SONA guidance, providing region-specific guidance. Extensive experiments demonstrate the effectiveness of our framework, achieving an impressive AUROC of 87% on near-OOD datasets, which surpasses the performance of baseline methods by a significant margin of approximately 6%.
BibTeX
@article{Yoon_Yoon_Sim_Choi_Lee_Cho_Lee_Lim_2025, title={Diffusion-based Semantic Outlier Generation via Nuisance Awareness for Out-of-Distribution Detection}, volume={39}, url={https://ojs.aaai.org/index.php/AAAI/article/view/33427}, DOI={10.1609/aaai.v39i12.33427}, abstractNote={Out-of-distribution (OOD) detection, determining whether a given sample is part of the in-distribution (ID) or not, has been newly explored by a generative model-based outlier synthesizing approach, especially with diffusion models. Nonetheless, existing diffusion models often produce outliers that are considerably distant from the ID in pixel-space, showing limited efficacy for capturing subtle distinctions between ID and OOD. To address these issues, we propose a novel framework, Semantic Outlier generation via Nuisance Awareness (SONA), which directly utilizes informative pixel-space ID images in diffusion models. Thereby, the generated outliers achieve two crucial properties: (i) they closely resemble the ID mainly in nuisances, while (ii) represent discriminative semantic information. To facilitate the separate effect on semantics and nuisances, we introduce SONA guidance, providing region-specific guidance. Extensive experiments demonstrate the effectiveness of our framework, achieving an impressive AUROC of 87% on near-OOD datasets, which surpasses the performance of baseline methods by a significant margin of approximately 6%.}, number={12}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Yoon, Suhee and Yoon, Sanghyu and Sim, Ye Seul and Choi, Sungik and Lee, Kyungeun and Cho, Hye-Seung and Lee, Hankook and Lim, Woohyung}, year={2025}, month={Apr.}, pages={13078-13086} }