B-scan ultrasonic image feature extraction of fatty liver based on gray level Co-occurrence matrix
ZHU Fu-zhen, WU Bin* (School of Information and Engineering, Southwest University of Science and Technology, Mianyang 621010, China)
Objective To provide a method of quantitative analysis in diagnosing fatty liver. Methods B-scan ultrasonic image of liver, extracts some features from B-scan ultrasonic image of fatty liver based on gray level Co-occurrence matrix was studied and these features were calculated and analyzed. Results and Conclusion The emulating result of these image processing is that these features, including angular second moment, entropy, and inversed differential moment, different prominently between normal liver and fatty liver, and can be used to distinguish B-scan ultrasonic image primly.