Nonlinear signal processing for vocal folds damage detection based on heterogeneous sensor network

Heterogeneous sensor network-based medical decision making could facilitate the patient diagnosis process. In this paper, we present an intelligent approach for vocal folds damage detection based on patient's vowel voices using heterogeneous sensor network. Based on human voice samples and Hidden Markov Model, we show that transformed voice samples (linearly combined samples) follow Gaussian distribution, further we demonstrate that a type-2 fuzzy membership function (MF), i.e., a Gaussian MF with uncertain mean, is most appropriate to model the transformed voices samples, which motivates us to use a nonlinear signal processing technique, interval type-2 fuzzy logic systems, to handle this problem. We also apply Short-Time-Fourier-Transform (STFT) and Singular-Value-Decomposition (SVD) to the vowel voice samples, and observe that the power decay rate could be used as an identifier in vocal folds damage detection. Two fuzzy classifiers, a Bayesian classifier and a linear classifier, are designed for vocal folds damage detection based on human vowel voices /a:/ and /i:/ only, and the fuzzy classifiers are compared against the Bayesian classifier and linear classifier. Simulation results show that an interval type-2 fuzzy classifier performs the best of the four classifiers. HighlightsHeterogeneous sensor network-based medical decision making could facilitate the patient diagnosis process.Two fuzzy classifiers, a Bayesian classifier and a linear classifier, are designed for vocal folds damage detection.Interval type-2 fuzzy classifier performs the best of the four classifiers.