跳到主要导航 跳到搜索 跳到主要内容

Glasses: Adapter-Enhanced DINOv3 with Boundary-Aware Multi-scale Representations for Medical Image Segmentation

  • Northwestern Polytechnical University Xian

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

摘要

Accurate medical image segmentation requires simultaneous modeling of robust global semantics and precise anatomical boundaries. While Vision Transformers (ViTs) exhibit strong global representation capability, they often struggle to capture fine-grained local structures and boundary details, particularly in scenarios with low contrast and blurred boundaries. To address this limitation, we propose Glasses, a novel framework that equips pre-trained Transformers with a structure-enhanced visual adaptation mechanism for medical image segmentation. Specifically, Glasses employs a frozen DINOv3 as the backbone encoder, with only the terminal layers fine-tuned. We introduce a Multi-Scale Deformable Attention Adapter that performs cross-scale interaction between lightweight convolutional structural priors and Transformer tokens extracted from specific intermediate layers of DINOv3, thereby transforming single-scale token embeddings into hierarchical representations enriched with both local structural cues and global context. This design preserves the stability of pre-trained semantic representations while improving the modeling of local anatomical structures and scale variations. Furthermore, we introduce an independent Edge-Aware Convolutional Encoder that explicitly learns boundary-enriched multi-scale representations under nested boundary supervision, providing a strong source of fine-grained local structural information that complements the adapter-enhanced Transformer features during decoding. A U-Net-style decoder then progressively fuses the adapter-enhanced Transformer features and the boundary-enriched convolutional representations to recover high-resolution segmentation outputs. Extensive experiments on the Synapse and ISIC datasets demonstrate that Glasses achieves Dice coefficients of 83.52% and 91.60%, respectively, significantly outperforming state-of-the-art (SOTA) methods. Source code is available at https://github.com/friay/Glasses.

源语言英语
主期刊名Medical Image Computing and Computer Assisted Intervention - MICCAI 2026 - 29th International Conference, 2026, Proceedings
编辑Guang Yang, Ehsan Adeli, Marleen de Bruijne, Marleen de Bruijne, Bartlomiej W. Papiez, Stefanie Speidel, Pallavi Tiwari, Guoyan Zheng, Mohammad Yaqub, Qi Dou, Islem Rekik
出版商Springer Science and Business Media Deutschland GmbH
页229-239
页数11
ISBN(印刷版)9783032380845
DOI
出版状态已出版 - 2027
活动29th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2026 - Strasbourg, 法国
期限: 27 9月 2026 → 1 10月 2026

丛书

姓名Lecture Notes in Computer Science
卷16884 LNCS
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议29th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2026
国家/地区法国
市Strasbourg
时期27/09/26 → 1/10/26

学术指纹

探究 'Glasses: Adapter-Enhanced DINOv3 with Boundary-Aware Multi-scale Representations for Medical Image Segmentation' 的科研主题。它们共同构成独一无二的学术指纹。

引用此