基于Mamba增强局部特征提取的皮肤病变分割模型

doi:10.16088/j.issn.1001-6600.2025122905

摘要/Abstract

摘要： 早期皮肤病变分割技术对皮肤癌的诊治起到十分重要的作用。然而,现有皮肤病变分割模型中,CNN难于提取图片的全局特征,Transformer模型计算开销大,导致训练成本居高不下,为此,本文提出LEM-UNet(local feature enhanced Mamba-UNet)模型。该模型基于vision Mamba-UNet重构其解码器,增强全局特征提取;提出多尺度特征提取残差模块RMS和局部特征提取模块LFM,加强模型对图像中不同尺寸大小病灶的分割能力,同时提升模型对病灶边界特征的捕获与提取精度,增强模型整体的分割效果。在ISIC2017和ISIC2018公共皮肤病变分割数据集上的对比实验表明,相较于基线模型,本文模型在mIoU、DSC、Acc、Spe、Sen评价指标上分别提升0.97、0.56、0.22、0.20、0.29和1.55、0.91、0.41、0.16、1.23个百分点,尤其在分割对象与背景对比度不高的场景下表现更为出色。

关键词: 皮肤病变分割, Mamba, 局部特征, 全局特征, 多尺度

Abstract: In the treatment of skin cancer, early skin lesion segmentation technology is of great importance. However, in existing skin lesion segmentation models, global features of images are difficult to be extracted by CNNs. Computational overhead of the model is increased by Transformers, resulting in persistently high training costs. To solve this problem, the Local Feature Enhanced Mamba U-Net (LEM-UNet) model is proposed in this paper. Based on Vision Mamba-UNet, the decoder is reconstructed to enhance global feature extraction. A multi-scale feature extraction residual module (RMS) and a local feature extraction module (LFM) are proposed. These modules strengthen the model’s ability to segment lesions of different sizes in images and improve the model’s accuracy in capturing and extracting lesion boundary features, thus enhancing the overall segmentation effect of the model. Comparative experiments are conducted on the ISIC2017 and ISIC2018 public skin lesion segmentation datasets. The results show that compared with the original baseline model, mIoU, DSC, Acc, Spe, and Sen increase by 0.97, 0.56, 0.22, 0.20, 0.29 percentage points and 1.55, 0.91, 0.41, 0.16, 1.23 percentage points respectively. The model performs particularly well in scenarios where the contrast between the segmented object and the background is not high.

Key words: skin lesion segmentation, Mamba, local features, global features, multi-scale

中图分类号: TP391.41

胡志强, 吕晓琪, 谷宇. 基于Mamba增强局部特征提取的皮肤病变分割模型[J]. 广西师范大学学报（自然科学版）, 2026, 44(5): 63-74.

Hu Zhiqiang, Lü Xiaoqi, Gu Yu. Skin lesion segmentation model based on improved Mamba local feature acquisition[J]. Journal of Guangxi Normal University(Natural Science Edition), 2026, 44(5): 63-74.

参考文献

[1] Rogers H W, Weinstock M A, Feldman S R, et al. Incidence estimate of nonmelanoma skin cancer (keratinocyte carcinomas) in the US population, 2012[J]. JAMA Dermatology, 2015, 151(10): 1081. DOI: 10.1001/jamadermatol.2015.1187.
[2] Siegel R L, Miller K D, Wagle N S, et al. Cancer statistics, 2023[J]. CA: A Cancer Journal for Clinicians, 2023, 73(1): 17-48. DOI: 10.3322/caac.21763.
[3] 林千里, 张文俊, 汪汇, 等. 皮肤黑色素瘤流行病学及防治研究进展[J]. 中国医药导报, 2019, 16(3): 28-32.
[4] 蒋清婷, 叶海良, 曹飞龙. 基于三路径网络的医学图像分割方法[J]. 模式识别与人工智能, 2024, 37(1): 1-12. DOI: 10.16451/j.cnki.issn1003-6059.202401001.
[5] 赵德春, 袁杨, 秦璐, 等. 改进型密集递归残差U-Net的皮肤病变图像分割[J]. 中国生物医学工程学报, 2025, 44(3): 291-300.
[6] 刘梦寒, 刘骊, 付晓东, 等. 特征学习和边界引导的皮肤病变图像分割网络[J/OL]. 计算机工程, 1-12[2025-12-29].https://doi.org/10.19678/j.issn.1000-3428.0070434.
[7] 肖宇庭, 吕晓琪, 谷宇, 等. 基于拆分残差网络的糖尿病视网膜病变分类[J]. 广西师范大学学报(自然科学版), 2024, 42(1): 91-101. DOI: 10.16088/j.issn.1001-6600.2023051805.
[8] Gu Y, Shi R, Yang S K, et al. Multi-disease X-ray image classification of the chest based on global and local fusion adaptive networks[J]. Current Medical Imaging Formerly Current Medical Imaging Reviews, 2024, 20: e15734056291283. DOI: 10.2174/0115734056291283240808045952.
[9] 孟祥福, 李佳讯, 俞纯林, 等. 深度学习在皮肤病变图像分割中的研究综述[J]. 计算机科学与探索, 2026, 20(1): 21-39.
[10] Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation[C]//Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015. Cham: Springer International Publishing Switzerland, 2015: 234-241. DOI: 10.1007/978-3-319-24574-4_28.
[11] Zhou Z W, Rahman Siddiquee M M, Tajbakhsh N, et al. UNet++: a nested U-Net architecture for medical image segmentation[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. Cham: SpringerNature Switzerland AG, 2018: 3-11. DOI: 10.1007/978-3-030-00889-5_1.
[12] Oktay O, Schlemper J, Le Folgoc L, et al. Attention U-Net: learning where to look for the pancreas[PP/OL]. V3. arXiv (2018-05-20)[2025-12-29].https://doi.org/10.48550/arXiv.1804.03999.
[13] 李大湘, 杨福杰, 刘颖, 等. 融入交叉注意力编码的皮肤病变分割网络[J]. 光学精密工程, 2024, 32(4): 609-621. DOI: 10.37188/OPE.20243204.0609.
[14] Lu Z K, She C Y, Wang W, et al. LM-Net: a light-weight and multi-scale network for medical image segmentation[J]. Computers in Biology and Medicine, 2024, 168: 107717. DOI: 10.1016/j.compbiomed.2023.107717.
[15] 王旭阳, 梁宇航. 多尺度非对称注意力遥感去雾Transformer[J]. 广西师范大学学报(自然科学版), 2026, 44(2): 77-89. DOI: 10.16088/j.issn.1001-6600.2025061001.
[16] Chen J N, Lu Y Y, Yu Q H, et al. TransUNet: transformers make strong encoders for medical image segmentation[PP/OL]. arXiv(2021-02-08)[2025-12-29].https://doi.org/10.48550/arXiv.2102.04306.
[17] Dosovitskiy A, Beyer L, Kolesnikov A, et al. An image is worth 16x16 words: transformers for image recognition at scale[PP/OL]. V2. arXiv(2021-06-03)[2025-12-29].https://doi.org/10.48550/arXiv.2010.11929.
[18] Cao H, Wang Y Y, Chen J, et al. Swin-unet: unet-like pure transformer forMedical image segmentation[C]//Computer Vision-ECCV 2022 Workshops. Cham: Springer, 2023: 205-218. DOI: 10.1007/978-3-031-25066-8_9.
[19] Gu A, Dao T. Mamba: Linear-time sequence modeling with selective state spaces[PP/OL]. V2. arXiv(2024-05-31)[2025-12-29]. https://doi.org.10.48550/arXiv.2312.00752.
[20] Zhu L H, Liao B C, Zhang Q, et al. Vision mamba: efficient visual representation learning with bidirectional state space model[PP/OL]. V3. arXiv (2024-11-14)[2025-12-29].https://doi.org/10.48550/arXiv.2401.09417.
[21] Jiao J B, Liu Y, Liu Y F, et al. VMamba: visual state space model[C]//Advances in Neural Information Processing Systems 37 (NeurIPS 2024). Red Hork, NY: Curran Associates Inc., 2024: 103031-103063. DOI: 10.52202/079017-3273.
[22] Ma J, Li F F, Wang B. U-mamba: enhancing long-range dependency for biomedical image segmentation[PP/OL]. arXiv(2024-01-09)[2025-12-29].https://doi.org/10.48550/arXiv.2401.04722.
[23] Ruan J C, Li J C, Xiang S C. VM-UNet: vision mamba UNet for medical image segmentation[J]. ACM Transactions on Multimedia Computing, Communications, and Applications, 2025: 3767748. DOI: 10.1145/3767748.
[24] Zhang M Y, Yu Y, Jin S, et al. VM-UNET-V2: rethinking vision mamba UNet forMedical image segmentation[C]//Bioinformatics Research and Applications. Singapore: Springer, 2024: 335-346. DOI: 10.1007/978-981-97-5128-0_27.
[25] Wu R K, Liu Y H, Ning G C, et al. UltraLight VM-UNet: parallel vision Mamba significantly reduces parameters for skin lesion segmentation[J]. Patterns, 2025, 6(11): 101298. DOI: 10.1016/j.patter.2025.101298.
[26] 贺蒙蒙, 张小艳, 李洪安. 基于Mamba结构的轻量级皮肤病变图像分割网络[J]. 图学学报, 2025, 46(6): 1257-1266. DOI: 10.11996/JG.j.2095-302X.2025061257.
[27] 王国凯, 张翔, 王顺芳. 多尺度和边界融合的皮肤病变区域分割网络[J]. 计算机科学与探索, 2024, 18(7): 1826-1837. DOI: 10.3778/j.issn.1673-9418.2306003.
[28] 余快, 宋宝贵, 邵攀, 等. 基于层级尺度交互的U-Net遥感影像建筑物提取方法[J]. 广西师范大学学报(自然科学版), 2025, 43(2): 121-132. DOI: 10.16088/j.issn.1001-6600.2024032002.
[29] Gu Y, Liu J Q, Yang L D, et al. SSANet: novel residual network for computer-aided diagnosis of pulmonary nodules in chest computed tomography[J]. International Journal of Imaging Systems and Technology, 2024, 34(5): e23176. DOI: 10.1002/ima.23176.
[30] Jha D, Riegler M A, Johansen D, et al. DoubleU-Net: a deep convolutional neural network for medical image segmentation[C]//2020 IEEE 33rd International Symposium on Computer-Based Medical Systems (CBMS). Piscataway, NJ: IEEE, 2020: 558-564. DOI: 10.1109/cbms49503.2020.00111.
[31] Zhang Y D, Liu H Y, Hu Q. TransFuse: fusing transformers and CNNs for medical image segmentation[C]//Medical Image Computing and Computer Assisted Intervention-MICCAI 2021. Cham: Springer, 2021: 14-24. DOI: 10.1007/978-3-030-87193-2_2.
[32] Peng Y P, Chen D Z, Sonka M. U-Net V2: rethinking the skip connections of U-Net for medical image segmentation[C]//2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI). Piscataway, NJ: IEEE, 2025: 1-5. DOI: 10.1109/ISBI60581.2025.10980742.
[33] Cheng Z H, Guo J T, Zhang J, et al. Mamba-sea: a mamba-based framework with global-to-local sequence augmentation for generalizable medical image segmentation[J]. IEEE Transactions on Medical Imaging, 2025, 44(9): 3741-3755. DOI: 10.1109/TMI.2025.3564765.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed