Analysis of abdominal fat tissue images acquired with continuously moving table MRI

The amount of visceral abdominal fat is correlated with the risk for hypertension and diabetes. In this work, an obesity study is monitored with continuously moving table whole body MRI equipped with surface coils and using the Dixon fat/water contrast and reconstruction. A method is presented for the repeatable and reliable differentiation of the imaged abdominal fat into subcutaneous and visceral. The images are first restored jointly for intensity uniformity and processed to segment the body abdominal region. Then, the exterior boundary of the visceral region separating it from the subcutaneous fat region is identified. The segmentations are performed with a combination of distance maps and graph cuts algorithms. The performance of the method has been demonstrated on data from 32 subjects.