Saved in:
Bibliographic Details
Main Authors: Xie, Yuxin, Zhou, Tao, Zhou, Yi, Chen, Geng
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.19364
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917786175930368
author Xie, Yuxin
Zhou, Tao
Zhou, Yi
Chen, Geng
author_facet Xie, Yuxin
Zhou, Tao
Zhou, Yi
Chen, Geng
contents Weakly-supervised medical image segmentation is a challenging task that aims to reduce the annotation cost while keep the segmentation performance. In this paper, we present a novel framework, SimTxtSeg, that leverages simple text cues to generate high-quality pseudo-labels and study the cross-modal fusion in training segmentation models, simultaneously. Our contribution consists of two key components: an effective Textual-to-Visual Cue Converter that produces visual prompts from text prompts on medical images, and a text-guided segmentation model with Text-Vision Hybrid Attention that fuses text and image features. We evaluate our framework on two medical image segmentation tasks: colonic polyp segmentation and MRI brain tumor segmentation, and achieve consistent state-of-the-art performance. Source code is available at: https://github.com/xyx1024/SimTxtSeg.
format Preprint
id arxiv_https___arxiv_org_abs_2406_19364
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SimTxtSeg: Weakly-Supervised Medical Image Segmentation with Simple Text Cues
Xie, Yuxin
Zhou, Tao
Zhou, Yi
Chen, Geng
Computer Vision and Pattern Recognition
Weakly-supervised medical image segmentation is a challenging task that aims to reduce the annotation cost while keep the segmentation performance. In this paper, we present a novel framework, SimTxtSeg, that leverages simple text cues to generate high-quality pseudo-labels and study the cross-modal fusion in training segmentation models, simultaneously. Our contribution consists of two key components: an effective Textual-to-Visual Cue Converter that produces visual prompts from text prompts on medical images, and a text-guided segmentation model with Text-Vision Hybrid Attention that fuses text and image features. We evaluate our framework on two medical image segmentation tasks: colonic polyp segmentation and MRI brain tumor segmentation, and achieve consistent state-of-the-art performance. Source code is available at: https://github.com/xyx1024/SimTxtSeg.
title SimTxtSeg: Weakly-Supervised Medical Image Segmentation with Simple Text Cues
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2406.19364