zobrazit českou verzi

Identification:
ZCU/KKY/2015/015



Year: 2015
Author: Psutka J.V.; Švec J.;Šmídl L.; Müller L.; Psutka J.

A prototype of the voice interactive system

The developed technology voice interactive system with a module of automatic speech recognition is able to search for the relevant information in large audiovisual archives, whether voice or written form of dialogue - the task domain was an archive of TV news programs. The system is able to recognize the user's speech and convert it into a text form and provide the user with the desired answer or make a required action (providing information, performing the desired action). The voice interactive system uses automatic speech recognition module (dictionary of more than one million words), the speech synthesis module and dialog module for intelligent interaction. The system can automatically handle new news programs. The audio input stream is recognized and indexed on verbal and phonetic level. We use modern technology Docker and document database MongoDB. Interactive dialogue system uses HTML5 web interface for the text input in combination with the WebRTC technology, which enables the voice interaction.