Semi-MedRef: Semi-Supervised Medical Referring Image Segmentation with Cross-Modal Alignment

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Yuchen, Zhao, Zhen, Liu, Yi, Zhou, Luping
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917499579138048
author Li, Yuchen
Zhao, Zhen
Liu, Yi
Zhou, Luping
author_facet Li, Yuchen
Zhao, Zhen
Liu, Yi
Zhou, Luping
contents Medical referring image segmentation (MRIS) requires pixel-level masks aligned with textual descriptions of anatomical locations, making annotation costly in low-label regimes. Semi-supervised learning (SSL) can mitigate this burden by leveraging unlabeled data, but its success hinges on maintaining reliable image-text alignment under perturbations. Most existing SSL-based referred segmentation methods use either independent or simplistic multi-modal perturbations (e.g., left-right flips), without fully addressing cross-modal alignment under strong augmentation, while CutMix, highly effective in single-modal SSL, remains underexplored in multi-modal settings due to its tendency to disrupt image-text coherence. We propose Semi-MedRef, a teacher-student SSL framework designed to explicitly maintain consistency between medical images and positional language through three alignment-preserving components: T-PatchMix, a cross-modal CutMix-style augmentation that synchronizes patch mixing with referring expressions via position-constrained and probability-driven rules; PosAug, a position-aware text augmentation that masks or fuzzes anatomical phrases; and ITCL, a position-guided image-text contrastive learning module, which leverages positional pseudo-labels to construct soft anatomical positives and strengthen medically grounded cross-modal alignment. Experiments on QaTa-COV19 and MosMedData+ demonstrate that Semi-MedRef consistently outperforms both fully supervised and semi-supervised baselines across all label regimes.
format Preprint
id arxiv_https___arxiv_org_abs_2605_15720
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Semi-MedRef: Semi-Supervised Medical Referring Image Segmentation with Cross-Modal Alignment
Li, Yuchen
Zhao, Zhen
Liu, Yi
Zhou, Luping
Computer Vision and Pattern Recognition
Machine Learning
Medical referring image segmentation (MRIS) requires pixel-level masks aligned with textual descriptions of anatomical locations, making annotation costly in low-label regimes. Semi-supervised learning (SSL) can mitigate this burden by leveraging unlabeled data, but its success hinges on maintaining reliable image-text alignment under perturbations. Most existing SSL-based referred segmentation methods use either independent or simplistic multi-modal perturbations (e.g., left-right flips), without fully addressing cross-modal alignment under strong augmentation, while CutMix, highly effective in single-modal SSL, remains underexplored in multi-modal settings due to its tendency to disrupt image-text coherence. We propose Semi-MedRef, a teacher-student SSL framework designed to explicitly maintain consistency between medical images and positional language through three alignment-preserving components: T-PatchMix, a cross-modal CutMix-style augmentation that synchronizes patch mixing with referring expressions via position-constrained and probability-driven rules; PosAug, a position-aware text augmentation that masks or fuzzes anatomical phrases; and ITCL, a position-guided image-text contrastive learning module, which leverages positional pseudo-labels to construct soft anatomical positives and strengthen medically grounded cross-modal alignment. Experiments on QaTa-COV19 and MosMedData+ demonstrate that Semi-MedRef consistently outperforms both fully supervised and semi-supervised baselines across all label regimes.
title Semi-MedRef: Semi-Supervised Medical Referring Image Segmentation with Cross-Modal Alignment
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2605.15720