Abstract:
The absence of medical resources at remote places prevents many patients from receiving a prompt and accurate diagnosis of cardiovascular disorders. To address this, we p...Show MoreMetadata
Abstract:
The absence of medical resources at remote places prevents many patients from receiving a prompt and accurate diagnosis of cardiovascular disorders. To address this, we proposed a novel deep learning model based on partially fine-tuned VGG16 and ConvMixer for the automatic identification of various heart valve diseases (HVDs) from phonocardiogram (PCG) signals. The method involves preprocessing PCG signals and converting them into gammatone filterbank (GFB) based 2-D time–frequency images. To generate time–frequency images, we used gammatonegram, gammatone cepstral coefficients (GTCCs), and gammatone discrete wavelet coefficients (GDWCs) techniques. These time–frequency images are augmented to reduce overfitting and then fed into a VGG16 model. Partial fine-tuning of the VGG16 model accelerates convergence and further improves performance. By extending the VGG16 model with ConvMixer, Global AveragePooling, Dense, and Softmax layers, we enhance its capacity to capture intricate patterns. The ConvMixer enriches spatial and channelwise features using Depthwise and Pointwise convolutions. We also performed an ablation analysis to highlight the effect of ConvMixer with VGG16. In addition, performance evaluation based on precision, recall, F1-score, test accuracy, and validation accuracy reveals the efficacy of the proposed method. Comparisons between gammatonegram, GTCC, and GDWC show superior performance of gammatonegram, achieving a test accuracy of 99.60% and validation accuracy of 99.75%. Our approach demonstrates significant advances over existing methods, offering a promising solution for remote diagnosis of HVDs.
Published in: IEEE Sensors Journal ( Volume: 25, Issue: 2, 15 January 2025)
Keywords assist with retrieval of results and provide a means to discovering other relevant content. Learn more.
- IEEE Keywords
- Index Terms
- Heart Valve ,
- Valvular Heart Disease ,
- VGG16 Model ,
- Phonocardiogram Signals ,
- Cardiovascular Disease ,
- Test Accuracy ,
- Dense Layer ,
- Validation Accuracy ,
- Softmax Layer ,
- Global Average Pooling ,
- Ablation Analysis ,
- Cepstral Coefficients ,
- Pointwise Convolution ,
- Spectroscopic ,
- Artificial Neural Network ,
- Data Augmentation ,
- Convolution Operation ,
- Mitral Valve ,
- Cardiac Cycle ,
- Image Signal ,
- Mitral Valve Prolapse ,
- Global Average Pooling Layer ,
- Depthwise Convolution ,
- Short-time Fourier Transform ,
- Heart Sound ,
- Mel-frequency Cepstral Coefficients ,
- N-channel ,
- Cardiac Signaling ,
- Discrete Cosine Transform ,
- Convolutional Block
- Author Keywords
Keywords assist with retrieval of results and provide a means to discovering other relevant content. Learn more.
- IEEE Keywords
- Index Terms
- Heart Valve ,
- Valvular Heart Disease ,
- VGG16 Model ,
- Phonocardiogram Signals ,
- Cardiovascular Disease ,
- Test Accuracy ,
- Dense Layer ,
- Validation Accuracy ,
- Softmax Layer ,
- Global Average Pooling ,
- Ablation Analysis ,
- Cepstral Coefficients ,
- Pointwise Convolution ,
- Spectroscopic ,
- Artificial Neural Network ,
- Data Augmentation ,
- Convolution Operation ,
- Mitral Valve ,
- Cardiac Cycle ,
- Image Signal ,
- Mitral Valve Prolapse ,
- Global Average Pooling Layer ,
- Depthwise Convolution ,
- Short-time Fourier Transform ,
- Heart Sound ,
- Mel-frequency Cepstral Coefficients ,
- N-channel ,
- Cardiac Signaling ,
- Discrete Cosine Transform ,
- Convolutional Block
- Author Keywords