Safe Image Generation via Lightweight Concept Erasure in Diffusion Models
IEEE Transactions on Image Processing · 2026 · 35 · 7898-7911
DOI: 10.1109/tip.2026.3712345 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Image Processing, 2026. Verfasst von Xiaoyu Geng und 4 weiteren Autor:innen. Die Arbeit nennt 58 Literaturangaben. Gefördert von: National Natural Science Foundation of China, National Research Foundation, Singapore, through the AI Singapore Program, Japan Science and Technology Agency (JST) and the Agency for Science, Technology and Research (A*STAR) through Japan–Singapore Joint Call.
Abstract (Originalsprache)
Text-to-image diffusion models have achieved remarkable progress in image synthesis, but their potential misuse for generating unauthorized or harmful content has raised growing safety concerns. This has created an urgent need for safe diffusion-based image generation methods that can selectively suppress sensitive concepts while preserving the model's general generative capability. Existing concept erasure approaches typically rely on either model fine-tuning or closed-form editing. However, they often suffer from two major limitations: (1 insufficient or excessive erasure, where the former fails to suppress target concepts and the latter disrupts benign semantics; and (2 degradation of non-target concepts, where removing target concepts undermines the generation of unrelated concepts, especially in multi-concept scenarios. To address these issues, we propose the Singular Value Eraser (SVEraser), a lightweight concept erasure module that removes specific concepts by optimizing singular-value offsets of weight matrices. Operating in a compact yet expressive singular-value space, SVEraser enables precise concept removal while reducing side effects on unrelated content. Moreover, once trained for different concepts, multiple SVErasers can be flexibly combined for multi-concept erasure. To further reduce interference, we introduce an eraser activation mechanism that adaptively selects the appropriate SVErasers during inference based on the input prompt. Extensive experiments on copyrighted objects, artistic styles, and explicit content demonstrate that our method achieves accurate target concept removal while preserving non-target semantics, providing a practical and reliable solution for safe diffusion-based image generation.
Förderung
- National Natural Science Foundation of China
- National Research Foundation, Singapore, through the AI Singapore Program
- Japan Science and Technology Agency (JST) and the Agency for Science, Technology and Research (A*STAR) through Japan–Singapore Joint Call
Zitieren
APA
Geng, X., Hui, S., Wang, Y., Zhou, J. T., & Wang, Z. (2026). Safe Image Generation via Lightweight Concept Erasure in Diffusion Models. IEEE Transactions on Image Processing, 35, 7898-7911. https://doi.org/10.1109/tip.2026.3712345
Vancouver
Geng X, Hui S, Wang Y, Zhou JT, Wang Z. Safe Image Generation via Lightweight Concept Erasure in Diffusion Models. IEEE Transactions on Image Processing. 2026;35:7898-7911. doi:10.1109/tip.2026.3712345
BibTeX
@article{Geng2026,
author = {Geng, Xiaoyu and Hui, Shuaixiong and Wang, Yuxin and Zhou, Joey Tianyi and Wang, Zheng},
title = {Safe Image Generation via Lightweight Concept Erasure in Diffusion Models},
journal = {IEEE Transactions on Image Processing},
year = {2026},
volume = {35},
pages = {7898--7911},
doi = {10.1109/tip.2026.3712345}
}
EUVIMED