An Automatic Segmentation and Quantitation Technique for Abdominal Fat from MR Images of Obese Rats

Bhanu Prakash KN¹, Venkatesh Gopalan¹, Swee Shean Lee¹, Sendhil S. Velan¹

Automatic segmentation and quantification of SAT and VAT from MR images were performed to study the influence of exercise and calorie restriction on obese rats. Distance regularized Level set for delineating the SAT, VAT regions, and fuzzy C-means for classification of fat, organs and non-fat regions was developed and implemented. T2W SE from 35 animals (pre- and post-interventions), L1 L5 of spine were acquired using 7T Bruker Clinscan. Results of segmentation and quantification showed significant decrease of SAT and VAT in exercise and calorie-restriction groups. The proposed method reduced processing time and eliminated inter-and intra- observer variability in quantification of fat.