Related Experiment Video
Updated: Sep 6, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.9K
CM-SegNet: A deep learning-based automatic segmentation approach for medical images by combining convolution and
Wenyu Xing1, Zhibin Zhu2, Dongni Hou3
1Center for Biomedical Engineering, School of Information Science and Technology, Fudan University, Shanghai, 200438, China; Human Phenome Institute, Fudan University, Shanghai, 200438, China.
Computers in Biology and Medicine
|July 3, 2022
Summary
We developed CM-SegNet, a deep learning model for accurate medical image segmentation. This faster, more precise automatic segmentation method shows great clinical potential.
Area of Science:
- Medical imaging analysis
- Deep learning for image segmentation
Background:
- Accurate lesion segmentation is crucial for clinical diagnosis but challenging due to low contrast.
- Manual segmentation is time-consuming, necessitating efficient automatic methods.
Purpose of the Study:
- To develop a deep learning-based automatic segmentation model (CM-SegNet) for medical images across different modalities.
- To improve the generalization performance and efficiency of medical image segmentation.
Main Methods:
- Proposed CM-SegNet, a deep learning model utilizing multiscale input and encoding-decoding architecture.
- Incorporated multilayer perceptron and convolution modules for enhanced feature extraction.
- Achieved communication across channels and spatial locations, considering inter-patch edge information.
Main Results:
- CM-SegNet demonstrated superior segmentation performance compared to previous methods across six diverse medical image datasets.
- The model achieved significantly shorter training times.
- Validated using 5-fold cross-validation, confirming its effectiveness on 3D medical images.
Conclusions:
- CM-SegNet offers a faster and more accurate solution for automatic medical image segmentation.
- The model exhibits strong potential for clinical applications in lesion detection and evaluation.
