•  
  •  
 

Turkish Journal of Electrical Engineering and Computer Sciences

Author ORCID Identifier

SINAM ASHINIKUMAR SINGH', 'Ashinikumar SINGH: 0009-0001-3873-1955

SINAM Ajitkumar SINGH', 'AJITKUMAR SINGH: 0000-0001-7148-1233

AHEIBAM DINAMANI SINGH', 'Dinamani SINGH: 0000-0001-6251-2595

DOI

10.55730/1300-0632.4087

Abstract

The analysis of heart sound signals constitutes a pivotal domain in healthcare, with the prediction of imbalanced heart sounds offering critical diagnostic insights. However, the inherent diversity in cardiac sound patterns presents a substantial challenge in predicting imbalanced signals. Many scientific disciplines have focused a great deal of emphasis on the problem of class inequality. We introduce an ensemble learning approach employing a convolutional neural network model-based deep learning algorithm to effectively tackle the challenges associated with predicting imbalanced heart sound signals. We use a Gammatone filter bank to extract relevant features from the heard sound signal. Our approach leverages a pre-trained convolutional neural network architecture, fine-tuning it with gammatonegram images to improve the classification performance. To overcome the challenges posed by imbalanced datasets, we integrate data augmentation into the image processing pipeline. The images are subsequently subjected to classification through deep convolutional neural network employing a transfer learning technique. This involves the utilization of convolutional neural network models such as AlexNet, SqueezeNet, GoogLeNet, and VGG19 to address concerns related to model overfitting. Our experimental results are rigorously validated using the publicly accessible PhysioNet 2016 dataset. The proposed ensemble methodology, incorporating AlexNet, SqueezeNet, and VGG19 models, demonstrated superior performance, attaining an accuracy of 99.51% and a sensitivity and specificity rate of 99.34% and 99.67% respectively. These results emphasize the substantial clinical promise inherent in our methodology, particularly in the realm of identifying imbalanced and noisy heart sound signals. This, in turn, serves to advance the diagnosis of cardiovascular diseases.

Keywords

Convolutional Neural Network, Deep Learning, Gammatonegram, Phonocardiogram, PhysioNet

First Page

555

Last Page

573

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Share

COinS