Several studies have compared the effects of use of totally speech based and totally text based interaction. [] concludes that participants’ repetitions of commands were more frequent when speech was used for both the user production and reception modes rather than text. This project aims at building an open-source voice based intelligent response system that can serve as a medical assistant for doctors practicing medicine in villages across the globe.
IVR - short for Interactive Voice Response - is the technology that automates the interaction with telephone caller. Historically, IVR solutions have used pre-recorded voice prompts and menus to present information and options to callers, and touch-tone telephone keypad entry (popularly known as DTMF) to gather caller response. However, these IVR solutions are very reliable in terms of their accuracy (partially due to it's static nature), they are highly inefficient while catering illiterate or semi-literate segment of the society. Hence, modern IVR solutions now enable input and responses that are gathered via spoken words with voice recognition. Although these systems are very far from being robust to noise in telephone channel or errors in voice recognition, they appear to be the key to bridge the information gap existing in illiterate section of the society.
An efficient E-Triage project would require the following components/functionalities in it to make it usable/popular among general population :
[] Ludovic Le Bigot, Jean-François Rouet, Eric Jamet, Effects of Speech- and Text-Based Interaction Modes in Natural Language Human-Computer Dialogue
[] Irina Kondratova, Multimodal Interaction for Mobile Learning