3D unsupervised anomaly detection through virtual multi-view projection and reconstruction: Clinical validation on low-dose chest computed tomography

Download PDF
Open Webpage

Kyungsu Kim, Seong Je Oh, Ju Hwan Lee, Myung Jin Chung

Published: 01 Jan 2024, Last Modified: 03 Oct 2024Expert Syst. Appl. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•We proposed VMPR-UAD for anomaly detection and localization in 3D chest low-dose CT.•VMPR-UAD utilized multi-view projection as a pre-processing technique.•VMPR-UAD enabled 3D visualization of anomalies through post-processing technique.•VMPR-UAD achieved an AUC of 0.963 on 3D chest low-dose CT scans.