Automatic echocardiographical feature extraction for left ventricular wall motion and volume changes visualization

This paper describes a new fully automatic fuzzy multi resolution-based system for cardiac left ventricular (LV) sequence analysis. The edge detection process utilized in the system uses multiscale spatial and temporal information in a fuzzy multiresolution framework to identify a single moving edge point for each one of the epicardial and endocardial boundaries for all frames in a cardiac cycle. The raw extracted boundaries are processed in the wavelet domain and finally scale-space filtering is used to define the close LV boundaries. The system is used to automatically extract ejection fractions and volume changes for real short axis echocardiographic data. The dynamics of the LV wall are also visualized.