Features Extracted from Long and Short Durations for 1 Heart Sound

Sik-Ho Tsang

·

Follow

3 min read

·

Mar 24, 2024

--

Listen

Share

A New Method for Heart Disease Detection: Long Short-Term Feature Extraction from Heart Sound Data
Long-Short Term Features (LSTF), by Gendarmerie and Coast Guard Academy, Kafkas University
2023 MDPI J. Sensors (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] [LSTM U-Net (LU-Net)] [DL Overview]
==== My Other Paper Readings Are Also Over Here ====

• Feature extraction is carried out using two different time scales. Short-term features are computed from five-second fragments of heart sound instances, whereas long-term features are extracted from the entire signal.
• The long-term features are combined with the short-term features to create a feature pool known as long short-term features for classification.

Outline

1. Long-Short Term Features (LSTF)
2. Results

1. Long-Short Term Features (LSTF)

1.1. Short-Term Features (STF)

• As shown above, the short-term features contain a total of 27 features, with 14 of them being extracted from the time, high-order statistics, energy, and frequency domains. The rest of the short-term features are Mel Frequency Cepstral Coefficients (MFFCs).

1.2. Long-Term Features (LTF)

1.3. Feature Pooling

• To merge the short- and long-term features, long-term features are added at the end of the short-term features. There are 27 short-term and 6 long-term features.
• Component analysis can be performed to reduce the feature sets, it is understood that all long-term features and 16 short-term features carry a significant amount of information.

1.4. Classification Models

• NB, SVM, kNN and Ensemble Methods are tried.

2. Results

LSTF without feature reduction obtains the best performance.