Voice Command Recognition with Dynamic Time Warping (DTW) using Graphics Processing Units (GPU) with Compute Unified Device Architecture (CUDA)

• Gustavo Poli UFSCar
• Joao F. Mari USP
• Jose Hiroki Saito UFSCar
• Alexandre L. M. Levada UFSCar

Resumo

Recently, we are attending to a huge evolution on the development of high performance computing platforms. Among these platforms, the GPU (Graphics Processing Units) stimulated by game industries, constantly demanding more graphical processing power, evolved from a simple graphical card to a general purpose computation parallel data processing device. This article shows the GPU's viability to general purpose computation, developing a speech recognition application inside. Dynamic Time Warping (DTW) is applied on a voice password identification. Normally, DTW requires large amount of data and processing time, so that it is an efficient technique to simple vocabulary, when the voice commands set is small. Using NVIDIA GeForce 8800 GTX, with 128 processing unit cores, and a CUDA (Compute Unified Device Architecture) software platform development architecture, the DTW application was implemented, and tested its performance.

Palavras-chave: Speech recognition, Graphics, Computer architecture, High performance computing, Computer industry, Toy industry, Concurrent computing, Data processing, Vocabulary, Software performance