Abstract
Alongside the currently used nasal swab testing, the COVID-19 pandemic situation would gain noticeable advantages from low-cost tests that are available at any-time, anywhere, at a large-scale, and with real time answers. A novel approach for COVID-19 assessment is adopted here, discriminating negative subjects versus positive or recovered subjects. The scope is to identify potential discriminating features, highlight mid and short-term effects of COVID on the voice and compare two custom algorithms. A pool of 310 subjects took part in the study; recordings were collected in a low-noise, controlled setting employing three different vocal tasks. Binary classifications followed, using two different custom algorithms. The first was based on the coupling of boosting and bagging, with an AdaBoost classifier using Random Forest learners. A feature selection process was employed for the training, identifying a subset of features acting as clinically relevant biomarkers. The other approach was centered on two custom CNN architectures applied to mel-Spectrograms, with a custom knowledge-based data augmentation. Performances, evaluated on an independent test set, were comparable: Adaboost and CNN differentiated COVID-19 positive from negative with accuracies of 100% and 95% respectively, and recovered from negative individuals with accuracies of 86.1% and 75% respectively. This study highlights the possibility to identify COVID-19 positive subjects, foreseeing a tool for on-site screening, while also considering recovered subjects and the effects of COVID-19 on the voice. The two proposed novel architectures allow for the identification of biomarkers and demonstrate the ongoing relevance of traditional ML versus deep learning in speech analysis.
【저자키워드】 deep learning, machine learning, Convolutional neural network, Random forest, support vector machine, nasal swab, CNNConvolutional neural network, MLmachine learning, SVMSupport vector machine, DLDeep Learning, MFCCMel-frequency Cepstral Coefficients, Mel-frequency Cepstral Coefficients, PPositive subjects, Positive subjects, RRecovered subjects, Recovered subjects, HHealthy control subjects, Healthy control subjects, NSNasal Swab, PCRPolymerase Chain Reaction-based molecular swabs, Polymerase Chain Reaction-based molecular swabs, 1EVowel /e/ vocal task, Vowel /e/ vocal task, 2SSentence vocal task, Sentence vocal task, 3CCough vocal task, Cough vocal task, PvsHPositive versus Healthy subjects comparison, Positive versus Healthy subjects comparison, RvsHRecovered versus Healthy subjects comparison, Recovered versus Healthy subjects comparison, CFSCorrelation-based Feature Selection, Correlation-based Feature Selection, RFRandom Forest, ReLuRectified Linear Unit, Rectified Linear Unit, ROCReceiver-Operating Curve, Receiver-Operating Curve, 【초록키워드】 COVID-19, Biomarker, Biomarkers, COVID-19 pandemic, COVID, CNN, Features, Accuracy, Analysis, binary, real time, Classifier, Algorithms, subject, individual, positive, forest, AdaBoost, Effect, approach, feature, independent, highlight, selection process, identify, collected, clinically, evaluated, applied, subjects, adopted, subset, acting, COVID-19 positive subjects, 【제목키워드】 Analysis, deep,