Saved in:
Bibliographic Details
Main Authors: Chen, Hao, Wang, Yiwei, Li, Songze
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.13039
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909964471107584
author Chen, Hao
Wang, Yiwei
Li, Songze
author_facet Chen, Hao
Wang, Yiwei
Li, Songze
contents Concept erasure, which fine-tunes diffusion models to remove undesired or harmful visual concepts, has become a mainstream approach to mitigating unsafe or illegal image generation in text-to-image models.However, existing removal methods typically adopt a unidirectional erasure strategy by either suppressing the target concept or reinforcing safe alternatives, making it difficult to achieve a balanced trade-off between concept removal and generation quality. To address this limitation, we propose a novel Bidirectional Image-Guided Concept Erasure (Bi-Erasing) framework that performs concept suppression and safety enhancement simultaneously. Specifically, based on the joint representation of text prompts and corresponding images, Bi-Erasing introduces two decoupled image branches: a negative branch responsible for suppressing harmful semantics and a positive branch providing visual guidance for safe alternatives. By jointly optimizing these complementary directions, our approach achieves a balance between erasure efficacy and generation usability. In addition, we apply mask-based filtering to the image branches to prevent interference from irrelevant content during the erasure process. Across extensive experiment evaluations, the proposed Bi-Erasing outperforms baseline methods in balancing concept removal effectiveness and visual fidelity.
format Preprint
id arxiv_https___arxiv_org_abs_2512_13039
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Bi-Erasing: A Bidirectional Framework for Concept Removal in Diffusion Models
Chen, Hao
Wang, Yiwei
Li, Songze
Computer Vision and Pattern Recognition
Cryptography and Security
Concept erasure, which fine-tunes diffusion models to remove undesired or harmful visual concepts, has become a mainstream approach to mitigating unsafe or illegal image generation in text-to-image models.However, existing removal methods typically adopt a unidirectional erasure strategy by either suppressing the target concept or reinforcing safe alternatives, making it difficult to achieve a balanced trade-off between concept removal and generation quality. To address this limitation, we propose a novel Bidirectional Image-Guided Concept Erasure (Bi-Erasing) framework that performs concept suppression and safety enhancement simultaneously. Specifically, based on the joint representation of text prompts and corresponding images, Bi-Erasing introduces two decoupled image branches: a negative branch responsible for suppressing harmful semantics and a positive branch providing visual guidance for safe alternatives. By jointly optimizing these complementary directions, our approach achieves a balance between erasure efficacy and generation usability. In addition, we apply mask-based filtering to the image branches to prevent interference from irrelevant content during the erasure process. Across extensive experiment evaluations, the proposed Bi-Erasing outperforms baseline methods in balancing concept removal effectiveness and visual fidelity.
title Bi-Erasing: A Bidirectional Framework for Concept Removal in Diffusion Models
topic Computer Vision and Pattern Recognition
Cryptography and Security
url https://arxiv.org/abs/2512.13039