Simple is what you need for efficient and accurate medical image segmentation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913895824752640 |
|---|---|
| author | Yu, Xiang Chen, Yayan He, Guannan Zeng, Qing Qin, Yue Liang, Meiling Luo, Dandan Liao, Yimei Ren, Zeyu Kang, Cheng Yang, Delong Liang, Bocheng Pu, Bin Yuan, Ying Li, Shengli |
| author_facet | Yu, Xiang Chen, Yayan He, Guannan Zeng, Qing Qin, Yue Liang, Meiling Luo, Dandan Liao, Yimei Ren, Zeyu Kang, Cheng Yang, Delong Liang, Bocheng Pu, Bin Yuan, Ying Li, Shengli |
| contents | While modern segmentation models often prioritize performance over practicality, we advocate a design philosophy prioritizing simplicity and efficiency, and attempted high performance segmentation model design. This paper presents SimpleUNet, a scalable ultra-lightweight medical image segmentation model with three key innovations: (1) A partial feature selection mechanism in skip connections for redundancy reduction while enhancing segmentation performance; (2) A fixed-width architecture that prevents exponential parameter growth across network stages; (3) An adaptive feature fusion module achieving enhanced representation with minimal computational overhead. With a record-breaking 16 KB parameter configuration, SimpleUNet outperforms LBUNet and other lightweight benchmarks across multiple public datasets. The 0.67 MB variant achieves superior efficiency (8.60 GFLOPs) and accuracy, attaining a mean DSC/IoU of 85.76%/75.60% on multi-center breast lesion datasets, surpassing both U-Net and TransUNet. Evaluations on skin lesion datasets (ISIC 2017/2018: mDice 84.86%/88.77%) and endoscopic polyp segmentation (KVASIR-SEG: 86.46%/76.48% mDice/mIoU) confirm consistent dominance over state-of-the-art models. This work demonstrates that extreme model compression need not compromise performance, providing new insights for efficient and accurate medical image segmentation. Codes can be found at https://github.com/Frankyu5666666/SimpleUNet. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_13415 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Simple is what you need for efficient and accurate medical image segmentation Yu, Xiang Chen, Yayan He, Guannan Zeng, Qing Qin, Yue Liang, Meiling Luo, Dandan Liao, Yimei Ren, Zeyu Kang, Cheng Yang, Delong Liang, Bocheng Pu, Bin Yuan, Ying Li, Shengli Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition I.4.6 While modern segmentation models often prioritize performance over practicality, we advocate a design philosophy prioritizing simplicity and efficiency, and attempted high performance segmentation model design. This paper presents SimpleUNet, a scalable ultra-lightweight medical image segmentation model with three key innovations: (1) A partial feature selection mechanism in skip connections for redundancy reduction while enhancing segmentation performance; (2) A fixed-width architecture that prevents exponential parameter growth across network stages; (3) An adaptive feature fusion module achieving enhanced representation with minimal computational overhead. With a record-breaking 16 KB parameter configuration, SimpleUNet outperforms LBUNet and other lightweight benchmarks across multiple public datasets. The 0.67 MB variant achieves superior efficiency (8.60 GFLOPs) and accuracy, attaining a mean DSC/IoU of 85.76%/75.60% on multi-center breast lesion datasets, surpassing both U-Net and TransUNet. Evaluations on skin lesion datasets (ISIC 2017/2018: mDice 84.86%/88.77%) and endoscopic polyp segmentation (KVASIR-SEG: 86.46%/76.48% mDice/mIoU) confirm consistent dominance over state-of-the-art models. This work demonstrates that extreme model compression need not compromise performance, providing new insights for efficient and accurate medical image segmentation. Codes can be found at https://github.com/Frankyu5666666/SimpleUNet. |
| title | Simple is what you need for efficient and accurate medical image segmentation |
| topic | Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition I.4.6 |
| url | https://arxiv.org/abs/2506.13415 |