MaskMed: Decoupled Mask and Class Prediction for Medical Image Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xie, Bin, Agam, Gady
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917092979113984
author Xie, Bin
Agam, Gady
author_facet Xie, Bin
Agam, Gady
contents Medical image segmentation typically adopts a point-wise convolutional segmentation head to predict dense labels, where each output channel is heuristically tied to a specific class. This rigid design limits both feature sharing and semantic generalization. In this work, we propose a unified decoupled segmentation head that separates multi-class prediction into class-agnostic mask prediction and class label prediction using shared object queries. Furthermore, we introduce a Full-Scale Aware Deformable Transformer module that enables low-resolution encoder features to attend across full-resolution encoder features via deformable attention, achieving memory-efficient and spatially aligned full-scale fusion. Our proposed method, named MaskMed, achieves state-of-the-art performance, surpassing nnUNet by +2.0% Dice on AMOS 2022 and +6.9% Dice on BTCV.
format Preprint
id arxiv_https___arxiv_org_abs_2511_15603
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MaskMed: Decoupled Mask and Class Prediction for Medical Image Segmentation
Xie, Bin
Agam, Gady
Computer Vision and Pattern Recognition
Medical image segmentation typically adopts a point-wise convolutional segmentation head to predict dense labels, where each output channel is heuristically tied to a specific class. This rigid design limits both feature sharing and semantic generalization. In this work, we propose a unified decoupled segmentation head that separates multi-class prediction into class-agnostic mask prediction and class label prediction using shared object queries. Furthermore, we introduce a Full-Scale Aware Deformable Transformer module that enables low-resolution encoder features to attend across full-resolution encoder features via deformable attention, achieving memory-efficient and spatially aligned full-scale fusion. Our proposed method, named MaskMed, achieves state-of-the-art performance, surpassing nnUNet by +2.0% Dice on AMOS 2022 and +6.9% Dice on BTCV.
title MaskMed: Decoupled Mask and Class Prediction for Medical Image Segmentation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.15603