AUTOMATIC ISOCONTOUR DETERMINATION FOR IODINE DOSE CALCULATION USING CELLULAR NEURAL NETWORKS ON SCINTIGRAPHIC IMAGES

Abstract

The hyperthyroid disease treatment consists in swallowing iodine 131. The quantity of these radio isotope results in an evaluation of the thyroid volume from a single scintigraphic image. In medical routine, the volume is calculated from a manual selection of an isocontour defining the boundary of the thyroid. We propose in this paper an automatic method to extract this boundary using Cellular Neural Network (CNN). Results show that our method is comparable to manual choice given by four experts. Studied on 35 patients with hyperthyroid diseases, we conclude that CNN is a comprising approach in segmentation of scintigraphic images