Related Experiment Videos
Fully automated, level set-based segmentation for knee MRIs using an adaptive force function and template: data from
Chunsoo Ahn1, Toan Duc Bui1, Yong-Woo Lee1
1School of Electronic and Electrical Engineering, Sungkyunkwan University, Suwon, South Korea.
Biomedical Engineering Online
|August 26, 2016
Summary
A new automated algorithm accurately segments all three knee cartilage tissues for diagnosing osteoarthritis (OA). This method improves segmentation accuracy over existing approaches, aiding in better OA diagnosis and management.
Area of Science:
- Biomedical Imaging
- Medical Image Analysis
- Osteoarthritis Research
Background:
- Osteoarthritis (OA) is a prevalent condition affecting knee cartilage.
- Accurate segmentation of knee cartilage is crucial for OA diagnosis.
- Existing segmentation methods lack full automation, sufficient accuracy, or include all three cartilage tissues.
Purpose of the Study:
- To develop a novel, fully automated algorithm for segmenting all three knee cartilage tissues.
- To improve the accuracy and completeness of knee cartilage segmentation for OA diagnosis.
Main Methods:
- A level set-based segmentation algorithm utilizing novel template data.
- Spatial fuzzy C-mean clustering for automatic contour initialization.
- Modified force function to enhance segmentation performance.
Main Results:
- Achieved Dice Similarity Coefficients (DSCs) of 87.1% (femoral), 84.8% (patellar), and 81.7% (tibial) cartilage.
- Demonstrated performance improvements of 8.8%, 4.3%, and 3.5% over existing methods for femoral, patellar, and tibial cartilage, respectively.
- Successfully applied to all three cartilage structures, unlike prior methods.
Conclusions:
- A novel, fully automated segmentation algorithm for three knee cartilage types was developed.
- The algorithm integrates a state-of-the-art level set approach with new knee template data.
- Experimental results indicate an average performance improvement of 5% compared to existing methods.