Agglomerating Large Vision Encoders via Distillation for VFSS Segmentation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zeng, Chengxi, Jiang, Yuxuan, Zhang, Fan, Gambaruto, Alberto, Burghardt, Tilo
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913775026700288
author Zeng, Chengxi
Jiang, Yuxuan
Zhang, Fan
Gambaruto, Alberto
Burghardt, Tilo
author_facet Zeng, Chengxi
Jiang, Yuxuan
Zhang, Fan
Gambaruto, Alberto
Burghardt, Tilo
contents The deployment of foundation models for medical imaging has demonstrated considerable success. However, their training overheads associated with downstream tasks remain substantial due to the size of the image encoders employed, and the inference complexity is also significantly high. Although lightweight variants have been obtained for these foundation models, their performance is constrained by their limited model capacity and suboptimal training strategies. In order to achieve an improved tradeoff between complexity and performance, we propose a new framework to improve the performance of low complexity models via knowledge distillation from multiple large medical foundation models (e.g., MedSAM, RAD-DINO, MedCLIP), each specializing in different vision tasks, with the goal to effectively bridge the performance gap for medical image segmentation tasks. The agglomerated model demonstrates superior generalization across 12 segmentation tasks, whereas specialized models require explicit training for each task. Our approach achieved an average performance gain of 2\% in Dice coefficient compared to simple distillation.
format Preprint
id arxiv_https___arxiv_org_abs_2504_02351
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Agglomerating Large Vision Encoders via Distillation for VFSS Segmentation
Zeng, Chengxi
Jiang, Yuxuan
Zhang, Fan
Gambaruto, Alberto
Burghardt, Tilo
Computer Vision and Pattern Recognition
Artificial Intelligence
The deployment of foundation models for medical imaging has demonstrated considerable success. However, their training overheads associated with downstream tasks remain substantial due to the size of the image encoders employed, and the inference complexity is also significantly high. Although lightweight variants have been obtained for these foundation models, their performance is constrained by their limited model capacity and suboptimal training strategies. In order to achieve an improved tradeoff between complexity and performance, we propose a new framework to improve the performance of low complexity models via knowledge distillation from multiple large medical foundation models (e.g., MedSAM, RAD-DINO, MedCLIP), each specializing in different vision tasks, with the goal to effectively bridge the performance gap for medical image segmentation tasks. The agglomerated model demonstrates superior generalization across 12 segmentation tasks, whereas specialized models require explicit training for each task. Our approach achieved an average performance gain of 2\% in Dice coefficient compared to simple distillation.
title Agglomerating Large Vision Encoders via Distillation for VFSS Segmentation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2504.02351