The folks over at McGill University have created an Automated Door Attendent, a multi-modal virtual secretary. Using the attendant, visitors may leave video messages, schedule appointments, or review web-based documents and demos. The purpose of the door attendant is to study the possibilities and limitations of artificial agents deployed in roles typically assigned to humans.

The Automated Door Attendant uses Sphinx-4 for speech recognition. They've written a paper Automated Door Attendant : Speech Integration that describes the process of developing and integrating speech into their system. Here's a quote:

The most attractive aspect of the Sphinx Speech Recognition System was it's high flexibility and modifiability in terms of it's front-end. Furthermore, in order to accomplish the tasks required of the speech side of ADA, modifications and customizations could be made without affecting the underlying architecture of the decoder.

I am quite impressed that these folks were able to select a speech recognition system, figure out how it works and integrate it into their large application, all in less than a month.


Post a Comment:
Comments are closed for this entry.

This blog copyright 2010 by plamere