Phase Recognition in Abdominal Contrast-Enhanced CT Scans

Accurately classifying the phases of contrast-enhanced abdominal computed tomography (CT) scans is crucial for the development of a fully automated system for interpreting these scans. However, current classification approaches based on 3D convolutional neural networks (CNNs) are computationally complex and have high latency, which can limit their practicality. This study aims to address this issue by developing and validating a fast and precise multi-phase classifier capable of recognizing the three main types of contrast phases in abdominal CT scans.