Fully Automatic Abdominal Fat Segmentation System from a Low Resolution CT Image

As the prevalence of obesity continues to rise with lifestyle changes in recent years, accurate tools for quantifying abdominal body and organ fat are critically needed to assist researchers investigating therapeutic and preventive measures against obesity and its comorbidities. Fatty infiltration of the liver, pancreas, and skeletal muscles are indicators of diabetes, the metabolic syndrome, and obesity. In this paper, it is important to be able to calculate the amount of abdominal fat for diagnosing cardiovascular disease or metabolism disease. Therefore, we propose a fully automatic abdominal fat segmentation system for quantifying abdominal fat, including subcutaneous adipose tissue and visceral adipose tissue. It can be divided into three parts. First, the abdominal region is assessed from a low resolution CT image. Next, the boundary line between subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) is found using our proposed methods. Finally, the amount of abdominal fat in the SAT or VAT regions is calculated. From experiment results, our proposed system has a higher than 90% successful rate for distinguishing the SAT and VAT regions based on a level consistency error and a detection rate index. And, it has average 82% successful rate from calculating artificial assessment evaluated by the two radiologists.