Analysis of nonstandardized stress echocardiography sequences using multiview dimensionality reduction

Mariana A. Nogueira, Mathieu De Craene, Sergio Sanchez-Martinez, Devyani Chowdhury, Bart H. Bijnens, Gemma Piella

Abstract: Highlights•Myocardial motion and heart rate data are obtained from handgrip exercise sequences.•Multiple Kernel Learning is used to obtain a compact representation of the data.•Each subject’s sequence can be seen as a low-dimensional trajectory.•Distinct healthy/pathological clusters of response can be identified.•Pathology-specific patterns can be retrieved from the low-dimensional space.