Article Details

Abstract

To address segmentation challenges of blurred boundaries and background noise in lumbar MRI images, this paper proposes DPS-UNet, a segmentation network based on receptive field amplification and path augmentation. Firstly, a 7-layer deep encoder is constructed to amplify the global receptive field, effectively capturing complex topological features of the lumbar spine and overcoming the limitations in receptive field of traditional networks. Secondly, a bottom-up path augmentation structure is introduced to efficiently transmit shallow spatial localization information to deep layers, enhancing the expressive power of feature pyramid. Simultaneously, a parameter-free SimAM attention module is embedded to adaptively suppress soft tissue noise and reinforce edge responses. Experimental results demonstrate that DPS-UNet significantly outperforms mainstream methods, with the HD95 distance decreasing from 7.21 to 2.78 and mIoU reaching 86.45%. The ablation experiments further validate the synergistic effectiveness of the deep encoder and feature enhancement strategies, indicating promising clinical application potential.

Keywords

lumbar image segmentation; U-Net; SimAM; path augmentation; receptive field amplification

References

1. [1] QU B，CAO J P，QIAN C，et al. Current Development and Prospects of Deep Learning in Spine Image Analysis: A Literature Review [J].Quantitative Imaging in Medicine and Surgery，2022，12（6）：3454-3479.
2. [2] PANG S M，PANG C L，ZHAO L，et al. SpineParseNet: Spine Parsing for Volumetric MR Image by a Two-stage Segmentation Framework with Semantic Image Representation [J].IEEE Transactions on Medical Imaging，2021，40（1）：262-273.
3. [3] LIU X B，SONG L P，LIU S，et al. A Review of Deep-learning-based Medical Image Segmentation Methods [J/OL].Sustainability，2021，13（3）：1224（2021-01-25）[2025-09-17].https://doi.org/10.3390/su13031224.
4. [4] CHEN J，LU Y，YU Q，et al. Transunet: Transformers Make Strong Encoders for Medical Image Segmentation [J/OL].arXiv:2102.04306 [cs.CV].（2021-02-08）.https://arxiv.org/abs/2102.04306.
5. [5] CAO H，WANG Y Y，CHEN J，et al. Swin-Unet: Unet-like Pure Transformer for Medical Image Segmentation [C]//Computer Vision-ECCV 2022 Workshops.Tel Aviv：Springer，2022：205-218.
6. [6] ZHOU Z W，SIDDIQUEE M M R，TAJBAKHSH N，et al. UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation [J].IEEE Transactions on Medical Imaging，2020，39（6）：1856-1867.
7. [7] HUANG H M，LIN L F，TONG R F，et al. UNet 3+: A Full-scale Connected UNet for Medical Image Segmentation [C]//ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing （ICASSP）.Barcelona：IEEE，2020：1055-1059.
8. [8] LIU Z，MAO H Z，WU C-Y，et al. A ConvNet for the 2020s [J/OL].arXiv:2201.03545 [cs.CV].（2022-01-10）.https://arxiv.org/abs/2201.03545.
9. [9] 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.Strasbourg：Springer，2021：14-24.
10. [10] YANG L X，ZHANG R-Y，LI L D，et al. SimAM: A Simple, Parameter-free Attention Module for Convolutional Neural Networks [C]//38th International Conference on Machine Learning (ICML).online：PMLR，2021：11863-11874.
11. [11] HOU Q B，ZHOU D Q，FENG J S. Coordinate Attention for Efficient Mobile Network Design [C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Nashville：IEEE，2021：13708-13717.
12. [12] LI L，QIN J，LYU L R，et al. ICUnet++: An Inception-CBAM Network Based on Unet++ for MR Spine Image Segmentation [J].International Journal of Machine Learning and Cybernetics，2023，14：3671-3683.

About the authors

He Zhiyuan (1997-), male, Han ethnicity, from Fuzhou, Jiangxi Province, is currently pursuing a master's degree. His research focuses on deep learning based lumbar spine image segmentation; Corresponding author: Wang Canhua (1978-), male, Han ethnicity, born in Jiujiang, Jiangxi Province, associate professor, expert in medical device evaluation and consulting in Jiangxi Province, director of the Traditional Chinese Medicine Health Big Data Branch of the China Association of Traditional Chinese Medicine Information, doctoral degree, research direction: health big data and artificial intelligence, biomedical engineering, and the application of machine learning in medical instruments.