Telephony-Based Voice Pathology Assessment Using Automated Speech Analysis

Abstract

A system for remotely detecting vocal fold pathologies using telephone-quality speech is presented. The system uses a linear classifier, processing measurements of pitch perturbation, amplitude perturbation and harmonic-to-noise ratio derived from digitized speech recordings. Voice recordings from the Disordered Voice Database Model 4337 system were used to develop and validate the system. Results show that while a sustained phonation, recorded in a controlled environment, can be classified as normal or pathologic with accuracy of 89.1%, telephone-quality speech can be classified as normal or pathologic with an accuracy of 74.2%, using the same scheme. Amplitude perturbation features prove most robust for telephone-quality speech. The pathologic recordings were then subcategorized into four groups, comprising normal, neuromuscular pathologic, physical pathologic and mixed (neuromuscular with physical) pathologic. A separate classifier was developed for classifying the normal group from each pathologic subcategory. Results show that neuromuscular disorders could be detected remotely with an accuracy of 87%, physical abnormalities with an accuracy of 78% and mixed pathology voice with an accuracy of 61%. This study highlights the real possibility for remote detection and diagnosis of voice pathology.

Effects of Audio Compression in Automatic Detection of Voice Pathologies

“…With respect to the SVM, the optimum working point has been searched for inside the grid C = [10 3 , 10 6 ] and γ = [10 −4 , 10 −2 ]. Regarding the GMM, the parameter M has been evaluated into M = [1,8]. The evaluation of the system has been carried out by means of a cross-validation strategy with 11 folds.…”

“…Recently, there has been some work on automatic detection of voice pathologies using audio files recorded under nonideal conditions [8]. The authors developed a system with voice registers transmitted over a telephone line.…”

A Transparent Decision Support Tool in Screening for Laryngeal Disorders Using Voice and Query Data

“…In this work, the utterance-based approach was used to compute the audio parameters. Information on how these parameters can be computed can be found in [12,25]. A set (of a similar size) of other voice parameters could probably provide higher classification accuracy.…”

“…Several measures computed from voice data are already widely used to quantify dysphonia changes and characterise outcomes of therapeutic and surgical treatment of laryngeal diseases [8][9][10][11]. It was demonstrated that even a voice signal, transmitted through a telephone channel, contains much information for the task of laryngeal pathology detection [12].…”

Análisis de la influencia de las técnicas de compresión de voz en la detección de anomalías vocales

“…En este último caso los autores solo han encontrado los siguientes trabajos: [15] propone un sistema para detectar de forma remota patologías de voz usando señales de voz con la calidad de un sistema de telefonía fija; los resultados de este trabajo muestran que mientras para señales de voz de alta calidad grabadas en ambientes controlados, la precisión en la detección de patologías de voz alcanza el 89,1%, la precisión alcanzada para señales de voz con calidad telefónica desciende al 74,2%. En [16] se reporta una precisión del 89,7% usando la metodología EMD (Empirical Mode Decomposition) para clasificación, asumiendo una relación señal-ruido del canal de 30 dB.…”