A deep neural network-based automatic mispronunciation detection in Bengali accented English speech

Puntos clave

• Automatic detection of mispronunciation can significantly improve speech recognition accuracy in Bengali accents.
• Key evidence shows that deep neural networks achieve an accuracy rate of over 90% in identifying mispronounced words.
• This analysis employs deep neural networks to train models on Bengali accented English speech data for effective detection.
• The findings support the need for advanced AI solutions in language processing tasks, enabling better communication in diverse settings.