Extraction of abdominal organ regions using three-dimensional mathematical morphology

Abstract

In this paper, a new approach to automatically extract organ regions from abdominal CT (computational tomography) data using three-dimensional (3D) mathematical morphology is proposed. Since morphological region extraction algorithm can directly obtain `shape' information, it is considered efficient for organ region segmentation and extraction by detecting their sizes and shapes. After a 3D morphological image processing system has been constructed, a differential top-hat transformation (DTT) algorithm is proposed for automatic extraction of organ regions. The proposed system was verified by an experiment showing that 95.3% of the organ regions were correctly extracted from 3D abdominal CT images.