CausalCLIPSeg: Unlocking CLIP's Potential in Referring Medical Image Segmentation with Causal Intervention

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Yaxiong, Wei, Minghong, Zheng, Zixuan, Hu, Jingliang, Shi, Yilei, Xiong, Shengwu, Zhu, Xiao Xiang, Mou, Lichao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908276029915136
author Chen, Yaxiong
Wei, Minghong
Zheng, Zixuan
Hu, Jingliang
Shi, Yilei
Xiong, Shengwu
Zhu, Xiao Xiang
Mou, Lichao
author_facet Chen, Yaxiong
Wei, Minghong
Zheng, Zixuan
Hu, Jingliang
Shi, Yilei
Xiong, Shengwu
Zhu, Xiao Xiang
Mou, Lichao
contents Referring medical image segmentation targets delineating lesions indicated by textual descriptions. Aligning visual and textual cues is challenging due to their distinct data properties. Inspired by large-scale pre-trained vision-language models, we propose CausalCLIPSeg, an end-to-end framework for referring medical image segmentation that leverages CLIP. Despite not being trained on medical data, we enforce CLIP's rich semantic space onto the medical domain by a tailored cross-modal decoding method to achieve text-to-pixel alignment. Furthermore, to mitigate confounding bias that may cause the model to learn spurious correlations instead of meaningful causal relationships, CausalCLIPSeg introduces a causal intervention module which self-annotates confounders and excavates causal features from inputs for segmentation judgments. We also devise an adversarial min-max game to optimize causal features while penalizing confounding ones. Extensive experiments demonstrate the state-of-the-art performance of our proposed method. Code is available at https://github.com/WUTCM-Lab/CausalCLIPSeg.
format Preprint
id arxiv_https___arxiv_org_abs_2503_15949
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CausalCLIPSeg: Unlocking CLIP's Potential in Referring Medical Image Segmentation with Causal Intervention
Chen, Yaxiong
Wei, Minghong
Zheng, Zixuan
Hu, Jingliang
Shi, Yilei
Xiong, Shengwu
Zhu, Xiao Xiang
Mou, Lichao
Computer Vision and Pattern Recognition
Referring medical image segmentation targets delineating lesions indicated by textual descriptions. Aligning visual and textual cues is challenging due to their distinct data properties. Inspired by large-scale pre-trained vision-language models, we propose CausalCLIPSeg, an end-to-end framework for referring medical image segmentation that leverages CLIP. Despite not being trained on medical data, we enforce CLIP's rich semantic space onto the medical domain by a tailored cross-modal decoding method to achieve text-to-pixel alignment. Furthermore, to mitigate confounding bias that may cause the model to learn spurious correlations instead of meaningful causal relationships, CausalCLIPSeg introduces a causal intervention module which self-annotates confounders and excavates causal features from inputs for segmentation judgments. We also devise an adversarial min-max game to optimize causal features while penalizing confounding ones. Extensive experiments demonstrate the state-of-the-art performance of our proposed method. Code is available at https://github.com/WUTCM-Lab/CausalCLIPSeg.
title CausalCLIPSeg: Unlocking CLIP's Potential in Referring Medical Image Segmentation with Causal Intervention
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.15949