Key result
A support vector machine utilizing Log Mel features achieved a 49.7% unweighted average recall on a three-category heart sound classification task using the Heart Sounds Shenzhen Corpus.
Why the study?
Machine listening can support cardiac auscultation for heart status monitoring, but relevant databases and benchmark studies remain scarce.
The introduction of the Heart Sounds Shenzhen Corpus provides a publicly accessible database and benchmark for machine learning approaches to heart sound recognition.
Modest 49.7% recall on three-class heart sound task cautions against clinical use; leaves open ML optimization using the new Shenzhen Corpus benchmark.
Auscultation of the heart is a widely studied technique, which requires precise hearing from practitioners as a means of distinguishing subtle differences in heart-beat rhythm. This technique is popular due to its non-invasive nature, and can be an early diagnosis aid for a range of cardiac conditions. Machine listening approaches can support this process, monitoring continuously and allowing for a representation of both mild and chronic heart conditions. Despite this potential, relevant databases and benchmark studies are scarce. In this paper, we introduce our publicly accessible database, the Heart Sounds Shenzhen Corpus (HSS), which was first released during the recent INTERSPEECH 2018 ComParE Heart Sound sub-challenge. Additionally, we provide a survey of machine learning work in the area of heart sound recognition, as well as a benchmark for HSS utilising standard acoustic features and machine learning models. At best our support vector machine with Log Mel features achieves 49.7% unweighted average recall on a three category task (normal, mild, moderate/severe).
No takes yet. Share an insight, caveat, or question.
Dong et al. (2019) studied Cardiovascular diseases (normal, mild, moderate/severe heart sounds) (n=170). Support Vector Machine (SVM) with Log Mel features vs. Recurrent Neural Network (RNN) models was evaluated on Unweighted average recall (UAR) for 3-class heart sound classification. A support vector machine utilizing Log Mel features achieved a 49.7% unweighted average recall on a three-category heart sound classification task using the Heart Sounds Shenzhen Corpus.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: