Brief Review — Cardiac Sound Classification Using Mel-Frequency Cepstral Coefficients (MFCC) and Artificial Neural Network (ANN)
MFCC+ANN
Cardiac Sound Classification Using Mel-Frequency Cepstral Coefficients (MFCC) and Artificial Neural Network (ANN)
MFCC+ANN, by Universitas Gadjah Mada Yogyakarta
2018 ICITISEE (Sik-Ho Tsang @ Medium)Heart Sound Classification
2013 … 2023 [2LSTM+3FC, 3CONV+2FC] [NRC-Net] [Log-MelSpectrum+Modified VGGNet] [CNN+BiGRU] [CWT+MFCC+DWT+CNN+MLP]
==== My Other Paper Readings Are Also Over Here ====
- MFCC is used as features extracted from heart sound.
- ANN is used as classification model with MFCC as input.
Outline
- MFCC+ANN
- Results
1. MFCC+ANN
1.1. MFCC
MFCC is extracted from the heart sound by the above stages such as FFT and DCT.
1.2. ANN
ANN with 1 hidden layer and the number of neurons as much as 17 and the activation function of logsig = 1/1 + exp (-n) also the number of epochs 1000, is used.
2. Results
Where in identifying 13 types of apex heart sounds reached 92% in previous studies because 11 types of heart sounds were classified correctly.
However, 2 types of abnormal heart sounds cannot be classified correctly, because both types have almost the same characteristic features.
