Implementation of accent recognition methods subsystem for eLearning systems

The results of the implementation of an external accent recognition system and its integration into massive open online courses platform Moodle are reported. Accent recognition becomes important in foreign languages learning to provide a feedback to a student on a presence of a certain unwanted accent in a foreign language pronunciation. Implementation of several accent recognition methods and their comparison is provided. It is shown that neural networks provide the most reliable recognition given the accented utterances from Wildcat Corpus of Native- and Foreign-Accented English.