RS-MSConvNet: A Novel End-to-End Pathological Voice Detection Model

Abstract

Recent studies have reported the success of multi-scale convolution neural network (MSConvNet) model for many classification applications due to its powerful ability of exploring multiscale convolution block to extract multi-scale representations to make a detection. However, a new design based on MSConvNet for pathological voice detection has not been explored. In this paper, we propose RS-MSConvNet, a novel end-to-end MSConvNet model using raw speech for pathological voice detection. The main contribution of the proposed RS-MSConvNet method is to exploit the multi-scale convolution block, followed by spatial-temporal feature block, and fully connected layer as classification. In addition, to further improve accuracy performance, we propose a novel hybrid detection model by integrating the feature extraction ability of the RS-MSConvNet model and the classifier of support vector machine (SVM) method, called RS-MSConvNet-SVM model. The effectiveness of our proposed models is investigated using the TORGO database. The experimental results reveal that the RS-MSConvNet model outperforms other baseline methods in the speaker-independent task. Moreover and as compared to the RS-MSConvNet-SVM model, a further improved accuracy is obtained using the RS-MSConvNet-SVM model. These outcomes exhibit that our proposed models are useful for pathological voice detection. INDEX TERMSPathological voice detection, end-to-end architecture, multi-scale convolution, spatialtemporal feature, hybrid model This article has been accepted for publication in IEEE Access.

Exploiting Data-Efficient Image Transformer-Based Transfer Learning for Valvular Heart Diseases Detection

“…In future work, we aim to investigate other attention mechanisms to further refine our proposed methods. We also plan to incorporate multi-scale convolutional neural networks [35] and phase information [49], [65] as supplementary data to enhance our methodologies. MFCCs based on EMD KNN 88.40 [5] MFCCs based on EMD MLP 88.80 [5] MFCCs based on EMD SVM 96.20 [5] MFCCs based on EMD DNN 98.90 [55] FDPCT VGG16-based transfer learning 94.00 [55] FDPCT ResNet-50-based transfer learning 98.00 [55] FDPCT Deep CNN 99.48 [30] Chirplet transform MCC 98.33 [47] log-mel spectrogram LSTM 95.00 [47] log-mel spectrogram CNN 99.67 [48] LPCC, MFCC SFF-HLSTM 99.…”

“…In ad-dition, the study examined various machine learning classifiers, including the support vector machine (SVM) [22]- [27], multiclass composite classifiers (MCC) [28], Multilayer Perceptron (MLP) [29], Random Forest (RF) [30], and k-Nearest Neighbor (k-NN) [31]- [34]. The effectiveness of hand-crafted feature extraction for classification is a significant part of these conventional approaches, demonstrating the need for expertise, e.g., speech processing tasks [35].…”