Automatic detection of hemorrhagic pericardial effusion on PMCT using deep learning - a feasibility study
- PDF / 986,168 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 94 Downloads / 225 Views
ORIGINAL ARTICLE
Automatic detection of hemorrhagic pericardial effusion on PMCT using deep learning - a feasibility study Lars C. Ebert 1 & Jakob Heimer 1 & Wolf Schweitzer 1 & Till Sieberth 1 & Anja Leipner 1 & Michael Thali 1 & Garyfalia Ampanozi 1
Accepted: 28 July 2017 / Published online: 18 August 2017 # Springer Science+Business Media, LLC 2017
Abstract Post mortem computed tomography (PMCT) can be used as a triage tool to better identify cases with a possibly non-natural cause of death, especially when high caseloads make it impossible to perform autopsies on all cases. Substantial data can be generated by modern medical scanners, especially in a forensic setting where the entire body is documented at high resolution. A solution for the resulting issues could be the use of deep learning techniques for automatic analysis of radiological images. In this article, we wanted to test the feasibility of such methods for forensic imaging by hypothesizing that deep learning methods can detect and segment a hemopericardium in PMCT. For deep learning image analysis software, we used the ViDi Suite 2.0. We retrospectively selected 28 cases with, and 24 cases without, hemopericardium. Based on these data, we trained two separate deep learning networks. The first one classified images into hemopericardium/not hemopericardium, and the second one segmented the blood content. We randomly selected 50% of the data for training and 50% for validation. This process was repeated 20 times. The best performing classification network classified all cases of hemopericardium from the validation images correctly with only a few false positives. The best performing segmentation network would tend to underestimate the amount of blood in the pericardium, which is the case for most networks. This is the first study that shows Electronic supplementary material The online version of this article (doi:10.1007/s12024-017-9906-1) contains supplementary material, which is available to authorized users. * Lars C. Ebert [email protected]
1
Institute of Forensic Medicine, University of Zurich, Winterthurerstrasse 190/52, 8057 Zurich, Switzerland
that deep learning has potential for automated image analysis of radiological images in forensic medicine. Keywords PMCT . Deep learning . Forensic imaging . Hemopericardium . Neural networks
Introduction In forensic medicine, identification of the deceased and determination of the cause and manner of death are the main objectives. Every pathologic finding, anomaly or foreign body must be thoroughly documented. This makes reading images in a forensic setting much more time consuming. For a single routine forensic case, image interpretation may take approximately 2 h, while in complex cases it can easily take longer. PMCT may be used as a triage tool to better identify cases in which non-natural causes of death are suspected. This may be especially relevant in personnel- or cost-critical situations [1]. Substantial data can be generated by modern medical scanners, especially in a forensic setting
Data Loading...
