Linear prediction and discrete wavelet transform to identify pathology in voice signals

An Experimental Analysis on Multicepstral Projection Representation Strategies for Dysphonia Detection

“…Specifically, the discrete wavelet transform (DWT) is traditionally used for making features based, for example, on the signal energy in the frequency domain, as performed by Tsanas et al [59] to detect the presence of Parkinson's by speech analysis. Similarly, the DWT was used to compose a representation proposed by Fonseca et al [60] that takes into account a version filtered by an inverse linear predictive filter (ILPF) of the voice signal to detect laryngeal infections from samples from a Brazilian database. Hammami et al [61] also used parameters obtained by DWT with the application of the concept of empirical mode decomposition (EMD), together with high-order statistics (HOS), to detect voice pathologies in a two-step classification scheme.…”

Acoustic investigation of speech pathologies based on the discriminative paraconsistent machine (DPM)

“…Much literature has recently appeared describing different techniques for SPD. On one hand, most of them, such as those documented in references [22,[24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43], are related to binary classifiers that distinguish between normal and pathologically-affected voices. On the other hand, a few articles present multi-class algorithms capable of identifying specific issues, such as those documented in references [15,[44][45][46], as follows.…”