VIKOOB: Visual context information for optimising ventilation therapy (BMWi)

VIKOOB is an application project of the north German Research Association for the development of an AI-Space for intelligent health systems (KI-SIGS), a winner of the innovation competition "Artificial Intelligence" of the Federal Ministry of Economics (BMWi). The creation of the KI-Space includes the development and setup of an AI platform, the definition and implementation of an R&D roadmap, and the AI education of the north German healthcare industry. As an application project, VIKOOB will demonstrate the advantages of using the AI-Space and actively participate in the development.
The goal of VIKOOB is the improved use of multimodal sensor data in ventilation therapy through the interaction of visually extracted context information (from multi-camera depth images), time series of vital signs and advanced AI algorithms (deep learning and probabilistic models). This will enable comprehensive patient assessments (e.g. regarding stress, body position etc.) and potentially new diagnostic/therapeutic applications.
The Institute of Medical Informatics (IMI) will especially advance the methodological development of AI algorithms (deep convolution networks for feature extraction from multi-camera depth images) for patient context analysis. The methods will be trained and evaluated on laboratory data as well as on clinical data sets (in cooperation with the Intensive Care Unit of the University Medical Center Hamburg-Eppendorf (UKE)). In cooperation with the Institute of Medical Electrical Engineering (IME) and the medical device manufacturer Dräger, a prototype for improved patient assessment/treatment in respiratory therapy will be developed.

BMWi project funding (2020-2023) 930'000€ (UzL: 264'000€ , IMI: 132'000€)

Selected Publications

  1. Hansen L., Siebert M., Diesel J., Heinrich M.P.
    Fusing Information from Multiple 2D Depth Cameras for 3D Human Pose Estimation in the Operating Room
    International Journal of Computer Assisted Radiology and Surgery (IJCARS, 2019)
     
  2. Bockelmann N., Graßhoff J., Hansen L., Bellani G., Heinrich M.P., Rostalski P.
    Deep Learning for Prediction of Diaphragm Activity from the Surface Electromyogram
    Current Directions in Biomedical Engineering (2019)

Project Team

Prof. Dr. Mattias Heinrich
Alexander Bigalke
M. Sc. Lasse Hansen