Skip to content. | Skip to navigation

Informatik 5
Information Systems
Prof. Dr. M. Jarke
Personal tools
You are here: Home Theses Semantic Anomaly Detection in Medical Time Series


Prof. Dr. M. Jarke
RWTH Aachen
Informatik 5
Ahornstr. 55
D-52056 Aachen
Tel +49/241/8021501
Fax +49/241/8022321

How to find us

Annual Reports





Semantic Anomaly Detection in Medical Time Series

Thesis type
  • Master
Student Sven Festag
Status Finished
Submitted in 2019
Proposal on 30. Jan 2019 11:00
Proposal room Library i5
Add proposal to calendar vCal

Medical bedside monitoring of patients relies on anomaly detection in sequence data. Especially the increase of data collected in this safety critical area expedites the need

for fast, accurate and automatic anomaly detection. The main goal of this project is to de fine and evaluate a system that can serve as a basis for surveillance software integrated into monitors that are used in intensive care. Thus, the system has to detect anomalous patterns that are specifi c to certain physiologic signals like the electrical activity of the heart (ECG), blood pressure etc.

Autoencoders based on recurrent neural networks with long short-term memory cells or gated recurrent units will be evaluated for this task. As labelled medical training data is a sparse resource, a clustering approach will be combined with these autoencoders. This  enables unsupervised training of the anomaly detection mechanism.

Document Actions