科研之友
微信
新浪微博
Facebook
分享链接
摘要
Osteoarthritis (OA) is a major global health issue due to aging populations. Infrared thermography provides functional information on thermal and vascular conditions of knee joints and can thus be used for knee OA screening. However, the thermal diagnostic procedure for various diseases often requires manual analysis and interpretation, which heavily depends on a clinician's personal experience. In this paper, an automated infrared thermographic analysis method for knee OA screening is developed based on the collected data of normal subjects and outpatients in clinics. 266 knee thermal images (166 normal, 100 abnormal) acquired in the China Rehabilitation Research Centre, Beijing, are used for the first trial. An effective knee feature extraction algorithm based on patella-centering is proposed. The extracted features are fed to a support vector machine (SVM) classifier to perform automated recognition. Experimental results indicate that the SVM classifier has an accuracy rate of 85.49%, a sensitivity of 85.72%, and a specificity of 85.51% in detecting normal and abnormal cases. The proposed automated system for knee thermal screening can thus provide quantitative reference information in assisting clinical diagnosis.
- 出版日期2013
- 单位清华大学; 中国科学院理化技术研究所; 首都医科大学
