Automatic Pronunciation Assistance on Video
María Pantoja · 2014
In this article we present a novel method that uses image and speech processing techniques to analyze video from a second language learner and provides speaker's pronunciation training. The pronunciation recommendation provided for the speaker will be selected from a database using machine learning techniques The model presented integrates speech and image recognition technology capable of quantizing and analyzing the learner's input providing feed-back data to evaluate the model's performance. The image/audio analysis and the expert system needed to provide recommendations to students are implemented on the GPU to allow for a fast feedback to students. Results show that our methodology accurately assigns pronunciation recommendations equivalent to those provided by a human second language (L2) instructor.