This paper describes a preliminary investigation of Voice Pathology Detection using Deep Neural Networks (DNN). We used voice recordings of sustained vowel /a/ produced at normal pitch from German corpus Saarbruecken Voice Database (SVD). This corpus contains voice recordings and electroglottograph signals of more than 2 000 speakers. The idea behind this experiment is the use of convolutional layers in combination with recurrent Long-Short-Term-Memory (LSTM) layers on raw audio signal. Each recording was split into 64 ms Hamming windowed segments with 30 ms overlap. Our trained model achieved 71.36% accuracy with 65.04% sensitivity and 77.67% specificity on 206 validation files and 68.08% accuracy with 66.75% sensitivity and 77.89% specificity on 874 testing files. This is a promising result in favor of this approach because it is comparable to similar previously published experiment that used different methodology. Further investigation is needed to achieve the state-of-the-art results.

This paper describes a method for Speech Emotion Recognition (SER) using Deep Neural Network (DNN) architecture with convolutional, pooling and fully connected layers. We used 3 class subset (angry, neutral, sad) of German Corpus (Berlin Database of Emotional Speech) containing 271 labeled recordings with total length of 783 seconds. Raw audio data were standardized so every audio file has zero mean and unit variance. Every file was split into 20 millisecond segments without overlap. We used Voice Activity Detection (VAD) algorithm to eliminate silent segments and divided all data into TRAIN (80%) VALIDATION (10%) and TESTING (10%) sets. DNN is optimized using Stochastic Gradient Descent. As input we used raw data without and feature selection. Our trained model achieved overall test accuracy of 96.97% on whole-file classification.

To diagnose and classify the dysarthric speech, speech language pathologist (SLP) conducts a listening test. On the basis of the scores given by listeners the dysarthria is diagnosed and assessed. The above mentioned method is costly, time consuming and not very accurate. Unlike the traditional method, this research proposes an automatic diagnosis and assessment of dysarthria. The aim of this paper is to diagnose and classify the severity of dysarthria. The speech disorder specific prosodic features are selected by using genetic algorithm. The diagnosis and assessment of dysarthric speech is done by support vector machines. During diagnosis the classification accuracy of 98% has been achieved. And 87% of the dysarthric speech utterances are correctly classified. The standard UASPEECH database has been used in this work.

The use of orthogonal projections on high-dimensional input and target data in learning frameworks is studied. First, we investigate the relations between two standard objectives in dimension reduction, preservation of variance and of pairwise relative distances. Investigations of their asymptotic correlation as well as numerical experiments show that a projection does usually not satisfy both objectives at once. In a standard classification problem, we determine projections on the input data that balance the objectives and compare subsequent results. Next, we extend our application of orthogonal projections to deep learning tasks and introduce a general framework of augmented target loss functions. These loss functions integrate additional information via transformations and projections of the target data. In two supervised learning problems, clinical image segmentation and music information classification, the application of our proposed augmented target loss functions increases the accuracy.