No abstract

Many children with speech sound disorders cannot pronounce the sibilant consonants correctly. We have developed a serious game, which is controlled by the children's voices in real time, with the purpose of helping children on practicing the production of European Portuguese (EP) sibilant consonants. For this, the game uses a sibilant consonant classifier. Since the game does not require any type of adult supervision, children can practice producing these sounds more often, which may lead to faster improvements of their speech. Recently, the use of deep neural networks has given considerable improvements in the classification of a variety of use cases, from image classification to speech and language processing. Here, we propose to use deep convolutional neural networks to classify sibilant phonemes of EP in our serious game for speech and language therapy. We compared the performance of several different artificial neural networks that used Mel frequency cepstral coefficients or log Mel filterbanks. Our best deep learning model achieves classification scores of 95.48% using a 2D convolutional model with log Mel filterbanks as input features. Such results are then further improved for specific classes with simple binary classifiers.

Children with fricative distortion errors have to learn how to correctly use the vocal folds, and which place of articulation to use in order to correctly produce the different fricatives. Here we propose a virtual tutor for fricatives distortion correction. This is a virtual tutor for speech and language therapy that helps children understand their fricative production errors and how to correctly use their speech organs. The virtual tutor uses log Mel filter banks and deep learning techniques with spectral-temporal convolutions of the data to classify the fricatives in children's speech by place of articulation and voicing. It achieves an accuracy of 90.40% for place of articulation and 90.93% for voicing with children's speech. Furthermore, this paper discusses a multidimensional advanced data analysis of the first layer convolutional kernel filters that validates the usefulness of performing the convolution on the log Mel filter bank.

Many children suffering from speech sound disorders cannot pronounce the sibilant consonants correctly. We have developed a serious game that is controlled by the children's voices in real time and that allows children to practice the European Portuguese sibilant consonants. For this, the game uses a sibilant consonant classifier. Since the game does not require any type of adult supervision, children can practice the production of these sounds more often, which may lead to faster improvements of their speech. Recently, the use of deep neural networks has given considerable improvements in classification for a variety of use cases, from image classification to speech and language processing. Here we propose to use deep convolutional neural networks to classify sibilant phonemes of European Portuguese in our serious game for speech and language therapy. We compared the performance of several different artificial neural networks that used Mel frequency cepstral coefficients or log Mel filterbanks. Our best deep learning model achieves classification scores of 95.48% using a 2D convolutional model with log Mel filterbanks as input features.

The distortion of sibilant sounds is a common type of speech sound disorder in European Portuguese speaking children. Speech and language pathologists (SLP) use different types of speech production tasks to assess these distortions. One of these tasks consists of the sustained production of isolated sibilants. Using these sound productions, SLPs usually rely on auditory perceptual evaluation to assess the sibilant distortions. Here we propose to use an isolated sibilant machine learning model to help SLPs assessing these distortions.Our model uses Mel frequency cepstral coefficients of the isolated sibilant phones and it was trained with data from 145 children. The analysis of the false negatives detected by the model can give insight into whether the child has a sibilant production distortion. We were able to confirm that there exist some relation between the model classification results and the distortion assessment of professional SLPs. Approximately 66% of the distortion cases identified by the model are confirmed by an SLP as having some sort of distortion or are perceived as being the production of a different sound.