Ljungvall, Ingrid
- Department of Clinical Sciences, Swedish University of Agricultural Sciences
Research article2022Peer reviewedOpen access
Bisgin, P.; Strube, T.; Henze, J.; Ljungvall, I.; Häggström, J.; Wess, G.; Stadler, J.; Schummer, C.; Meister, S.; Howar, F.M.
Auscultation methods enable non-invasive diagnosis of diseases, e.g. of the heart, based on heartbeat sounds. Regular, early examinations using machine learning techniques could help to detect diseases at an early stage to prevent serious health conditions and then provide optimal therapy through continuous monitoring. There is already a lot of work on human data using AI algorithms to detect patterns in signals or images. However, there is hardly no work on detecting heart murmurs with digital such as Myxomatous Mitral Valve Disease. In this paper, we present a canine auscultation project that aims to provide a tool to establish a baseline of classification parameters from audio signals that could be used to monitor canine health status by analyzing deviations from this baseline. In the future, data analysis could also lead to prediction and early detection of other diseases.
AI; Classification; Digital Biomarkers; Machine-Learning; MMVD; Pattern Recognition
Current Directions in Biomedical Engineering
2022, Volume: 8, number: 2, pages: 765-768 Publisher: Walter de Gruyter GmbH
Clinical Science
DOI: https://doi.org/10.1515/cdbme-2022-1195
https://res.slu.se/id/publ/129689