Automatic disordered sound repetition recognition in continuous speech using CWT and kohonen network

Abstract

Automatic disorders recognition in speech can be very helpful for a therapist while monitoring therapy progress of patients with disordered speech. This article is focused on sound repetitions. The signal is analyzed using Continuous Wavelet Transform with 16 bark scales. Using the silence finding algorithm, only speech fragments are automatically found and cut. Each cut fragment is converted into a fixed-length vector and passed into the Kohonen network. Finally, the Kohonen winning neuron result is put on the 3-layer perceptron. Most of the analysis was performed and the results were obtained using the authors' program WaveBlaster. We use the STATISTICA package for finding the best perceptron which was then imported back into WaveBlaster and used for automatic blockades finding. The problem presented in this article is a part of our research work aimed at creating an automatic disordered speech recognition system.

Prolongation Recognition in Disordered Speech Using CWT and Kohonen Network

“…Quite large recognition statistics was created obtaining very high recognition ratios. Most of the theoretical aspects of this work are exactly the same as in our previous article [5], because in both cases we describe smaller parts of the one, bigger project. Therefore in chapters 2 and 3 we place only brief description of this theory (more details are in our previous article [5]).…”

“…Therefore 0-th neuron always pulls silence (which is always the weakest signal) to the top-left corner, then top-left corner (with neighbors) gathers weak signal, therefore strong signal is naturally placed in bottom-right corner. We also added additional step into the learning process [4] which was not used in our previous research [5]. This step is applied after the network has been trained using the standard algorithm described previously.…”

The assessment of speech fluency

“…Interesującą propozycją wydaje się "Wave Blaster" -wszechstronne narzędzie do analizy mowy przy użyciu ciągłej transformaty falkowej ze skalami barkowymi [29]. Program ten bazuje na specjalnie opracowanych algorytmach detekcji trzech rodzajów niepłynności: przedłużeń dźwięków, powtórzeń dźwięków oraz powtórzeń sylab i umożliwia wieloparametrowy ogląd danych w postaci: oscylogramów, spektrogramów, widm czy obrysów sieci Kohonena [30][31][32]. Należy jednak zaznaczyć, że wspomniane narzędzie diagnostyczne jest wciąż w fazie optymalizacji do zastosowania w praktyce klinicznej.…”