Segmentation-based heart sound feature extraction combined with classifier models for a VSD diagnosis system

作者：

Highlights：

• A diagnosis system is proposed for diagnosing a VSD.

• The diagnostic features [FG,FW] based on a novel method are extracted.

• A classification boundary method based on the SVM is proposed.

• Ellipse fitting is employed to build ellipse models for boundary curves.

• The performance is evaluated by clinical heart disease.

摘要

•A diagnosis system is proposed for diagnosing a VSD.•The diagnostic features [FG,FW] based on a novel method are extracted.•A classification boundary method based on the SVM is proposed.•Ellipse fitting is employed to build ellipse models for boundary curves.•The performance is evaluated by clinical heart disease.

论文关键词：VSD,Heart sound segmentation,Classifier models,SVM,Envelopes ET and EF,MWHT

论文评审过程：Available online 11 September 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.08.076