Histogram-based quantitative grading on ultrasonographic images of a fatty liver

Abstract

An abnormal fat accumulation in the liver can damage the liver cells and cause liver diseases. If the fat accumulation in the liver is diagnosed in the early stages; the functional abnormalities associated with fatty liver can be revealed and the severity of its can be assessed. There are several medical imaging techniques to diagnose fatty liver and determine the grade of fat in the liver. One of these techniques is the ultrasound imaging, which is non-invasive, cost-effective and easily accessible. However, there is always a need for an expert opinion to determine the level of fat in the liver from ultrasound images. Furthermore, the level of fat in the same liver image may be interpreted differently by the same or another expert when evaluated at different times. In order to increase objectivity and assist the expert, in this paper, a new quantitative measurement method called the Weighted Histogram (WH) is proposed to determine and grade the fatty liver from ultrasound images. The proposed WH method was compared with the results of the Gray Relational analysis method used for the same purpose in the literature. A Graphical User Interface (GUI) is designed to grade the fatty liver with MATLAB to improve ease of use. The proposed methodology will not only reduce false diagnosis caused by subjective differences in clinical assessment, but also quantitative assessment of fatty liver and grade by early diagnosis.

Keywords

• fatty liver
• liver fat level
• ultrasound image processing
• weighted histogram
• matlab gui