Med-Tuning: A New Parameter-Efficient Tuning Framework for Medical Volumetric Segmentation

Jiachen Shen, Wenxuan Wang, Chen Chen, Jianbo Jiao, Jing Liu, Yan Zhang, Shanshan Song, Jiangyun Li
Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, PMLR 250:1412-1433, 2024.

Abstract

The p̈re-training then fine-tuning (FT)\"{paradigm} is widely adopted to boost the model performance of deep learning-based methods for medical volumetric segmentation. However, conventional full FT incurs high computational and memory costs. Thus, it is of increasing importance to fine-tune pre-trained models for medical volumetric segmentation tasks in a both effective and parameter-efficient manner. In this paper, we introduce a new framework named Med-Tuning to realize parameter-efficient tuning (PET) for medical volumetric segmentation task and an efficient plug-and-play module named Med-Adapter for task-specific feature extraction. With a small number of tuned parameters, our framework enhances the 2D baselinesś precision on segmentation tasks, which are pre-trained on natural images. Extensive experiments on three benchmark datasets (CT and MRI modalities) show that our method achieves better results than previous PET methods on volumetric segmentation tasks. Compared to full FT, Med-Tuning reduces the fine-tuned model parameters by up to 4x, with even better segmentation performance. Our project webpage is at \url{https://rubics-xuan.github.io/Med-Tuning/}.

Cite this Paper

BibTeX
@InProceedings{pmlr-v250-shen24a, title = {Med-Tuning: A New Parameter-Efficient Tuning Framework for Medical Volumetric Segmentation}, author = {Shen, Jiachen and Wang, Wenxuan and Chen, Chen and Jiao, Jianbo and Liu, Jing and Zhang, Yan and Song, Shanshan and Li, Jiangyun}, booktitle = {Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning}, pages = {1412--1433}, year = {2024}, editor = {Burgos, Ninon and Petitjean, Caroline and Vakalopoulou, Maria and Christodoulidis, Stergios and Coupe, Pierrick and Delingette, Hervé and Lartizien, Carole and Mateus, Diana}, volume = {250}, series = {Proceedings of Machine Learning Research}, month = {03--05 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v250/main/assets/shen24a/shen24a.pdf}, url = {https://proceedings.mlr.press/v250/shen24a.html}, abstract = {The p̈re-training then fine-tuning (FT)\"{paradigm} is widely adopted to boost the model performance of deep learning-based methods for medical volumetric segmentation. However, conventional full FT incurs high computational and memory costs. Thus, it is of increasing importance to fine-tune pre-trained models for medical volumetric segmentation tasks in a both effective and parameter-efficient manner. In this paper, we introduce a new framework named Med-Tuning to realize parameter-efficient tuning (PET) for medical volumetric segmentation task and an efficient plug-and-play module named Med-Adapter for task-specific feature extraction. With a small number of tuned parameters, our framework enhances the 2D baselinesś precision on segmentation tasks, which are pre-trained on natural images. Extensive experiments on three benchmark datasets (CT and MRI modalities) show that our method achieves better results than previous PET methods on volumetric segmentation tasks. Compared to full FT, Med-Tuning reduces the fine-tuned model parameters by up to 4x, with even better segmentation performance. Our project webpage is at \url{https://rubics-xuan.github.io/Med-Tuning/}.} }
Endnote
%0 Conference Paper %T Med-Tuning: A New Parameter-Efficient Tuning Framework for Medical Volumetric Segmentation %A Jiachen Shen %A Wenxuan Wang %A Chen Chen %A Jianbo Jiao %A Jing Liu %A Yan Zhang %A Shanshan Song %A Jiangyun Li %B Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ninon Burgos %E Caroline Petitjean %E Maria Vakalopoulou %E Stergios Christodoulidis %E Pierrick Coupe %E Hervé Delingette %E Carole Lartizien %E Diana Mateus %F pmlr-v250-shen24a %I PMLR %P 1412--1433 %U https://proceedings.mlr.press/v250/shen24a.html %V 250 %X The p̈re-training then fine-tuning (FT)\"{paradigm} is widely adopted to boost the model performance of deep learning-based methods for medical volumetric segmentation. However, conventional full FT incurs high computational and memory costs. Thus, it is of increasing importance to fine-tune pre-trained models for medical volumetric segmentation tasks in a both effective and parameter-efficient manner. In this paper, we introduce a new framework named Med-Tuning to realize parameter-efficient tuning (PET) for medical volumetric segmentation task and an efficient plug-and-play module named Med-Adapter for task-specific feature extraction. With a small number of tuned parameters, our framework enhances the 2D baselinesś precision on segmentation tasks, which are pre-trained on natural images. Extensive experiments on three benchmark datasets (CT and MRI modalities) show that our method achieves better results than previous PET methods on volumetric segmentation tasks. Compared to full FT, Med-Tuning reduces the fine-tuned model parameters by up to 4x, with even better segmentation performance. Our project webpage is at \url{https://rubics-xuan.github.io/Med-Tuning/}.
APA
Shen, J., Wang, W., Chen, C., Jiao, J., Liu, J., Zhang, Y., Song, S. & Li, J.. (2024). Med-Tuning: A New Parameter-Efficient Tuning Framework for Medical Volumetric Segmentation. Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 250:1412-1433 Available from https://proceedings.mlr.press/v250/shen24a.html.

Related Material