Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 2, 2026Franklin OpenOpen Access

Phonocardiogram classification Based on Machine learning and Deep learning using hybrid features extraction techniques

View Full Paper
Ask AI
Bookmark
Share

Key result

Combining FBC and DWT features improves PCG classification to ~98% accuracy for detecting heart disorders.

Why the study?

Cardiovascular ailments require early and accurate diagnosis, which can be aided by analyzing phonocardiogram signals using signal processing and machine learning.

Population

Database of PCG signals comprising one normal and three pathological cardiac sound categories

Comparison

Amalgamated FBC and DWT features vs individual feature extraction across SVM, NB, KNN, and GRU models

Authors

IJIrfan JavidSHShamaila HayatUniversity of Poonch RawalakotTBTuba BatoolUniversity of Poonch Rawalakot

Discussion

Loading...

Member takes

Implication

Promising for AI-assisted heart disorder screening by clinicians; extends PCG machine learning but should not yet change practice.

Key Points

  • The aim is to develop a reliable method for classifying cardiac sounds using phonocardiogram signals and machine learning techniques.
  • Utilized phonocardiogram signals for analysis of normal and abnormal cardiac sounds.
  • Employed Discrete Wavelet Transform (DWT) and Filter Bank Coefficients (FBC) for feature extraction.
  • Applied four machine learning models: SVM, Naïve Bayes, KNN, and GRU for classification.
  • Combined FBC and DWT features to enhance model training and classification accuracy.
  • Achieved classification accuracy of up to 97.8% for detecting heart disorders.
  • Combination of extracted features significantly improved outcomes across all model evaluations.

Structured PICO

P
Population
Database comprising four distinct cardiac sound signal categories: one normal and three pathological, derived from multiple sources.
I
Intervention
Machine learning and deep learning models (SVM, NB, KNN, GRU) trained on amalgamated features extracted using Discrete Wavelet Transform (DWT) and Filter Bank Coefficients (FBC).
O
Outcome
Classification accuracy of heart disorders

Combining Filter Bank Coefficients and Discrete Wavelet Transform feature extraction techniques with machine learning models enables highly accurate classification of phonocardiogram signals for diagnosing heart disorders.

Cite This Study

Javid et al. (2026) studied this question. Combining Filter Bank Coefficients and Discrete Wavelet Transform features improved PCG classification accuracy, achieving up to 97.8% in detecting heart disorders.

synapsesocial.com/papers/69a528ecf1e85e5c73bf04f1https://doi.org/10.1016/j.fraope.2026.100554
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1A lightweight hybrid deep learning system for cardiac valvular disease classification2022 · 83 citations
  2. 2Best subsequence selection of heart sound recording based on degree of sound periodicity2011 · 26 citations
  3. 3Wavelets for computer graphics: a primer. 21995 · 159 citations
  4. 4Robust Biometric Recognition From Palm Depth Images for Gloved Hands2015 · 113 citations