B-scan Image Feature Extraction of Fatty Liver

Fatty liver is a common phenomenon concerned in today's healthcare. Using B-ultrasound modality can help identify the abnormal disease in cardiology. This paper presents an extraction method for fatty liver ultrasound image feature. Through the extraction, one can easily analyzes ultrasound images from clinical practices in fatty liver pathology. After selecting a region of interest in a B-ultrasonic image, and introducing a median filter, the quality of image is significantly improved. The selected image region is divided into several small regional areas. By changing the selected image gray-scale and making the average gray of these small pieces, the image features can be easily identified. The difference between abnormal and normal liver are finally identifies through multi-grade gray levels contraction. Experimental results from the proposed method demonstrated its feasibility and applicability for high performance fatty liver classification.