Related Experiment Video
Updated: Jul 11, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.8K
MFF-Net: Multiscale feature fusion semantic segmentation network for intracranial surgical instruments
Zhenzhong Liu1,2, Laiwang Zheng1,2, Shubin Yang1,2
1Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin, China.
Summary
This study introduces an efficient network for segmenting surgical instruments in craniotomy environments, achieving high accuracy and speed. The model offers a valuable reference for advancing intelligent surgical robots.
Area of Science:
- Medical Robotics
- Computer Vision
- Surgical Technology
Background:
- Automatic segmentation of surgical instruments is vital for safety in robot-assisted surgery.
- Challenges in craniotomy environments include occlusion and illumination variations.
- An efficient surgical instrument segmentation network is proposed to address these issues.
Purpose of the Study:
- To develop an efficient surgical instrument segmentation network for craniotomy environments.
- To overcome challenges like occlusion and illumination in surgical instrument image analysis.
- To enhance surgical safety through accurate automated segmentation.
Main Methods:
- Utilizes YOLOv8 as the object detection framework with an integrated semantic segmentation head.
- Employs concatenation of multi-channel feature maps to fuse deep and shallow features for improved generalization.
- Incorporates the GBC2f module for network lightweighting and global information capture.
Main Results:
- Achieved a Mean Average Precision (MPA) score of 94.9% on an intracranial glioma surgical instrument dataset.
- Obtained a Mean Intersection over Union (MIoU) value of 89.9%.
- Demonstrated a processing speed of 126.6 Frames Per Second (FPS).
Conclusions:
- The proposed segmentation model significantly outperforms existing state-of-the-art models.
- Experimental results validate the model's effectiveness in complex surgical scenarios.
- This research provides a valuable reference for the development of intelligent surgical robots.
