CRCE: Coreference-Retention Concept Erasure in Text-to-Image Diffusion Models


Yuyang Xue (University of Edinburgh), Edward Moroshko (University of Edinburgh), Feng Chen (University of Edinburgh), Jingyu Sun (University of Edinburgh), Steven G. McDonagh (University of Edinburgh), Sotos Tsaftaris (University of Edinburgh)
The 35th British Machine Vision Conference
PDFPosterVideo (Right click to download)Supplementary

Abstract

Text-to-Image diffusion models can produce undesirable content that necessitates concept erasure. However, existing methods struggle with under-erasure, leaving residual traces of targeted concepts, or over-erasure, mistakenly eliminating unrelated but visually similar concepts. To address these limitations, we introduce CRCE, a novel concept erasure framework that leverages Large Language Models to identify semantically related concepts that should be erased alongside the target, in addition to distinct concepts that should be preserved. By explicitly modelling coreferential and retained concepts semantically, CRCE enables more precise concept removal, without unintended erasure. Additionally, we contribute CorefConcept, a comprehensive dataset encompassing objects, intellectual property, and personal identities, which we make publicly available to support future research in concept erasure. The code and dataset are available online.

Citation

@inproceedings{Xue_2025_BMVC,
author    = {Yuyang Xue and Edward Moroshko and Feng Chen and Jingyu Sun and Steven G. McDonagh and Sotos Tsaftaris},
title     = {CRCE: Coreference-Retention Concept Erasure in Text-to-Image Diffusion Models},
booktitle = {36th British Machine Vision Conference 2025, {BMVC} 2025, Sheffield, UK, November 24-27, 2025},
publisher = {BMVA},
year      = {2025},
url       = {https://bmva-archive.org.uk/bmvc/2025/assets/papers/Paper_88/paper.pdf}
}


Copyright © 2025 The British Machine Vision Association and Society for Pattern Recognition
The British Machine Vision Conference is organised by The British Machine Vision Association and Society for Pattern Recognition. The Association is a Company limited by guarantee, No.2543446, and a non-profit-making body, registered in England and Wales as Charity No.1002307 (Registered Office: Dept. of Computer Science, Durham University, South Road, Durham, DH1 3LE, UK).

Imprint | Data Protection