FUZZY CLUSTERING ON ABDOMINAL MRI FOR ADIPOSE TISSUE QUANTIFICATION

摘要

In this paper, we present an automatic and reliable segmentation to quantify intra-abdominal to subcutaneous adipose tissue ratio, in metabolic syndrome diagnosis. We use dual echo Magnetic Resonance (MR) images acquired in phase and out of phase. The segmentation is provided by two fuzzy algorithms: fuzzy generalized clustering and fuzzy connectedness algorithm. A result validation is performed with a comparison of expert results with purpose-made software results.