Related Experiment Video
Updated: Oct 5, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.0K
Liver segmentation in CT imaging with enhanced mask region-based convolutional neural networks
Xiaowen Chen1, Xiaoqin Wei1, Mingyue Tang2
1School of Medical Imaging, North Sichuan Medical College, Nanchong, China.
Annals of Translational Medicine
|January 24, 2022
Summary
This study introduces an improved Mask R-CNN method for accurate liver segmentation in CT scans. The enhanced algorithm achieves superior performance in identifying liver characteristics for disease diagnosis.
Area of Science:
- Medical Imaging
- Computer Vision
- Artificial Intelligence
Background:
- Liver segmentation in CT imaging is vital for disease diagnosis but challenging due to poor contrast, image noise, and anatomical variations.
- Existing methods struggle with accuracy and robustness in complex abdominal scans.
Purpose of the Study:
- To develop a novel and robust liver segmentation method for CT image sequences.
- To enhance the accuracy of liver segmentation by addressing challenges like poor contrast and image noise.
Main Methods:
- An enhanced Mask R-CNN model integrated with graph-cut segmentation was developed.
- The k-nearest neighbor (k-NN) algorithm was used for pixel clustering and aspect ratio determination.
- Rotation-invariant object recognition and a fully convolutional network (FCN) were employed for precise localization and segmentation.
Main Results:
- The proposed Mask R-CNN algorithm demonstrated superior performance compared to conventional methods.
- Key metrics such as the Dice Similarity Coefficient (DSC) and MICCAI metrics showed significant improvements.
Conclusions:
- The improved Mask R-CNN architecture offers high performance, accuracy, and robustness for liver segmentation in CT imaging.
- This method provides a reliable tool for analyzing liver characteristics and aiding in the diagnosis of liver diseases.
