Carotid artery image segmentation using modified spatial fuzzy c-means and ensemble clustering
Mehdi Hassan1, Asmatullah Chaudhry, Asifullah Khan
1Department of Computer & Information Sciences, PIEAS, P.O. Nilore, Islamabad, Pakistan.
Computer Methods and Programs in Biomedicine
|September 18, 2012
Summary
This study introduces an improved image segmentation technique for carotid artery ultrasound images to detect plaque. The enhanced method improves accuracy and speed, aiding in disease diagnosis.
Area of Science:
- Medical imaging
- Biomedical engineering
- Artificial intelligence in medicine
Background:
- Ultrasound imaging is crucial for non-invasive disease diagnosis but suffers from low image quality due to noise and interference.
- Accurate diagnosis from carotid artery ultrasound images is challenging, necessitating advanced image processing techniques like segmentation.
- Existing segmentation methods may not fully leverage spatial information, impacting diagnostic efficiency.
Purpose of the Study:
- To develop an improved image segmentation technique for enhanced plaque detection in carotid artery ultrasound images.
- To improve the accuracy and efficiency of disease diagnosis from carotid artery ultrasound imaging.
- To utilize ensemble clustering and feature selection for robust carotid artery plaque identification.
Main Methods:
- An improved spatial fuzzy c-means and ensemble clustering approach was developed for image segmentation.
- Spatial, wavelet, and Gray Level Co-occurrence Matrix (GLCM) features were extracted and optimized using a genetic search.
- Intima-media thickness (IMT) was measured, and Multi-Layer Back-Propagation Neural Networks (MLBPNN) were used for classification.
Main Results:
- The proposed ensemble clustering with a reduced feature set demonstrated superior segmentation time and clustering accuracy.
- The MLBPNN classifier exhibited effective learning capabilities for classifying images as normal or abnormal based on IMT.
- The technique successfully identified plaque presence in carotid artery ultrasound images.
Conclusions:
- The developed image segmentation and classification approach is effective for detecting carotid artery plaque.
- The method offers a valuable tool for improving the accuracy and efficiency of cardiovascular disease diagnosis.
- This research highlights the potential of advanced AI techniques in medical image analysis for clinical applications.
