Laryngeal pathology detection by means of class-specific neural maps

Abstract

Most of the existing systems and methods for laryngeal pathology detection are characterized by a classification error. One of the basic problems is the approximation and estimation of the probability density functions of the given classes. In order to increase the accuracy of laryngeal pathology detection and to eliminate the most dangerous error--classification of a patient with laryngeal disease as a normal speaker--here an approach based on modeling of the probability density functions (pdf's) of the input vectors of the normal and pathological speakers by means of two prototype distribution maps (PDM), respectively, is proposed. The pdf of the input vectors of an unknown normal or pathological speaker is also modeled by such a prototype distribution neural map--PDM(X)--and the pathology detection is done by means of a ratio of specific similarities rather than by a direct comparison of some type of distance/similarity with a threshold. The experiments show an increased classification accuracy and that the proposed method can be used for screening the laryngeal diseases. The method is applied in a consulting system for clinical practice.

Glottal Source biometrical signature for voice pathology detection

“…Description x 1 Pitch x 2 Jitter x [3][4][5] 3 different estimations of shimmer x [6][7] parameters related with glottal closure x [8][9][10] 3 parameters related with HNR x [11][12][13][14] MWC power spectral density in 4 bins x [15][16][17][18][19][20][21][22][23] Amplitude of the MWC PSD singularities as described in (14) x [24][25][26][27][28][29][30][31][32] Position of the MWC PSD singularities as described in (14) x [33][34] Slenderness of the first and second "V" notches as described in (14) x [35][36][37] Estimations of the vocal fold body biomechanical parameters as described in (15) x [38][39][40] Estimations of the vocal fold cover biomechanical parameters as described in (15) x [41][42][43] Vocal fold body biomechanical parameter unbalance as described in (15) x [44][45]<...>…”

“…Typical perturbation parameters are jitter, shimmer and harmonics-to-noise ratios in their different interpretations. This has been the traditional approach and is well documented in the works of different researchers [1][2] [3][4] [5]. The main objection to this methodology is that the original voice signal is contaminated by phonetic-acoustic information related to specific articulation features, and therefore perturbation parameters derived from this signal are influenced by articulation, making it difficult to grant new advances towards pathology classification.…”

The Effectiveness of the Glottal to Noise Excitation Ratio for the Screening of Voice Disorders

“…To date, very few studies have evaluated the discriminative capabilities of the acoustic parameters. With the same database used in this study, Parsa evaluated the discriminative capabilities of several noise features 17 [17][18][19][20][21] indicate that an accurate screening can be carried out using a combination of several of the above-mentioned acoustic parameters, enabling each individual voice utterance to be quantified by a single set of one-dimensional parameters (similar to those enumerated above). Although the multidimensional studies reported a good efficiency in screening (obtaining accuracies up to 96%), 20,21 such analysis is not always easy to interpret from the perspective of a human evaluator and it is usually carried out by methods based on complex pattern recognition techniques.…”

Automatic detection of voice impairments from text-dependent running speech

“…Such parameters (jitter, shimmer, Harmonics to Noise Ratio (HNR), Normalized Noise Energy (NNE), Voice Turbulence Index (VTI), Glottal to Noise Excitation Ratio (GNE), Signal to Noise Ratio (SNR), Frequency Amplitude Tremor (FATR), and many others [1][2][3][4][5][6]) were developed to measure quality and ''degree of normality'' of voice registers from the sustained phonation of vowels. Using this acoustic material, previous studies [7][8][9] indicate that the detection of voice alterations can be carried out by means of the above-mentioned long-term averaged acoustic parameters, enabling each individual voice utterance to be quantified by a single vector. The main drawback of most of these parameters is that they relay on an accurate estimation of the fundamental frequency, a rather complicated task in presence of certain pathologies.…”