# Recognizing Abnormal Heart Sounds Using Deep Learning

arXiv: https://arxiv.org/abs/1707.04642 · HTML: https://mlsanity.com/paper/1707.04642/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Jul 2017.

## Built on by (most cited first, 9 of 9)

- [Deep Learning for Time Series Anomaly Detection: A Survey](https://mlsanity.com/paper/2211.05244/) (2022)
- [Machine learning based disease diagnosis: A comprehensive review](https://mlsanity.com/paper/2112.15538/) (2021)
- [Designing ECG Monitoring Healthcare System with Federated Transfer Learning and Explainable AI](https://mlsanity.com/paper/2105.12497/) (2021)
- [A Comprehensive Survey on Heart Sound Analysis in the Deep Learning Era](https://mlsanity.com/paper/2301.09362/) (2023)
- [Diagrammatization and Abduction to Improve AI Interpretability With Domain-Aligned Explanations for Medical Diagnosis](https://mlsanity.com/paper/2302.01241/) (2023)
- [The Significance of Machine Learning in Clinical Disease Diagnosis: A Review](https://mlsanity.com/paper/2310.16978/) (2023)
- [Exploring traditional machine learning for identification of pathological auscultations](https://mlsanity.com/paper/2209.00672/) (2022)
- [Comparison of window shapes and lengths in short-time feature extraction for classification of heart sound signals](https://mlsanity.com/paper/2604.13567/) (2026)
- [Comparing Spectrogram Front-Ends for Abnormal Heart-Sound Detection with a Convolutional Neural Network](https://mlsanity.com/paper/2607.16220/) (2026)

## Abstract (arXiv)

The work presented here applies deep learning to the task of automated cardiac auscultation, i.e. recognizing abnormalities in heart sounds. We describe an automated heart sound classification algorithm that combines the use of time-frequency heat map representations with a deep convolutional neural network (CNN). Given the cost-sensitive nature of misclassification, our CNN architecture is trained using a modified loss function that directly optimizes the trade-off between sensitivity and specificity. We evaluated our algorithm at the 2016 PhysioNet Computing in Cardiology challenge where the objective was to accurately classify normal and abnormal heart sounds from single, short, potentially noisy recordings. Our entry to the challenge achieved a final specificity of 0.95, sensitivity of 0.73 and overall score of 0.84. We achieved the greatest specificity score out of all challenge entries and, using just a single CNN, our algorithm differed in overall score by only 0.02 compared to the top place finisher, which used an ensemble approach.