Concept for Automatic Multi-object Organ Detection and Segmentation in Abdominal CT Data

Abstract

The ability to generate 3D patient models in a fast and reliable way, is of great importance, e.g. for the simulation of liver punctures in a virtual reality simulation. The aim is to automatically detect and segment abdominal structures in CT scans. In particular among the selected organ group, the pancreas poses a challenge. We use a combination of random regression forests and U-Nets to detect bounding boxes and generate segmentation masks for five abdominal organs (liver, kidneys, spleen, and pancreas). Proof of concept training and testing was carried out on 50 CT scans from various public sources. Preliminary results showed Dice coefficients of up to 0.71. The proposed method can theoretically be used for any anatomical structure, as long as sufficient training data is available.