Persian Accents Identification Using an Adaptive Neural Network
Azam Rabiee, Saeed Setayeshi · 2010
Speaker's accent can reduce the performance of automatic speech recognition systems. This paper considered Persian accent identification using a model includes preprocessing, feature extraction and neural networks. Samples are from five different Persian accents. The performance of the neural networks as an adaptive approach is compared with two statistical approaches. The effect of increasing the number of accents in performance is shown here too.