Related Experiment Video
Updated: Jun 9, 2025
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
370
MCI Net: Mamba- Convolutional lightweight self-attention medical image segmentation network.
Yelin Zhang1, Guanglei Wang1,2, Pengchong Ma1
1College of Electronic and Information Engineering, Hebei University, Hebei 071002, People's Republic of China.
Biomedical Physics & Engineering Express
|October 24, 2024
Summary
A new lightweight deep learning model, MCI-Net, offers efficient and accurate medical image segmentation. It reduces parameters and complexity for faster lesion identification, aiding timely diagnosis and treatment.
Area of Science:
- Medical image analysis
- Deep learning for healthcare
- Computer-aided diagnosis
Background:
- Current deep learning models for medical image segmentation face challenges with complex variations and irregular shapes, leading to issues like incomplete information extraction, large model sizes, and high computational demands.
- Existing networks struggle to efficiently process intricate medical image data, impacting diagnostic speed and subsequent treatment planning.
Purpose of the Study:
- To introduce MCI-Net, a novel lightweight network designed for efficient and accurate medical image segmentation.
- To address the limitations of existing models by reducing parameter count and computational complexity while maintaining high performance.
Main Methods:
- MCI-Net employs linear modeling on sequences to effectively mark features and filter irrelevant information.
- The network efficiently captures local-global information using a reduced number of channels and attention mechanisms with exchange value mapping.
- Model lightweighting is achieved through these methods, enabling thorough local-global information interaction and establishing semantic relationships.
Main Results:
- MCI-Net demonstrates superior performance compared to other advanced networks across five public datasets: X-ray, Lung, ISIC-2016, ISIC-2018, and capsule endoscopy/gastrointestinal segmentation.
- Ablation studies on four datasets further validate the effectiveness of the proposed MCI-Net architecture.
- The network boasts a low parameter count (5.48 M), computational complexity (4.41), and time complexity (0.263).
Conclusions:
- MCI-Net offers a lightweight, accurate, and high-performance solution for medical image segmentation.
- The proposed model significantly reduces diagnosis time, providing a valuable reference for clinical applications.
- This research contributes to the advancement of efficient deep learning models in medical imaging.
