TV Content Search Using Voice
As the availability of television content continues to grow, a more effective method to search for programs is needed. We have developed a system that allows television viewers to search an electronic program guide (EPG) by voice. Our proposed solution is to add a microphone to remote controls to enable voice input for searching over variable collections of programs available through EPGs.
Background & Objective: Consumers find television programs from a growing list of diverse sources including broadcast, cable, satellite, video-on-demand, Internet and disks. The need for effective search of content is growing as the number of choices for TV viewing and/or recording explodes. Existing search methods, such as text entry using conventional TV remote controls is awkward. This project aims to provide a more effective search using voice input.
Technical Discussion: Our solution adds a microphone to remote controls that allows the user to quickly find programs or browse the nearest matches using voice. The system utilizes MERL's SpokenQuery technology in a new domain of EPG search. SpokenQuery allows search terms to be entered that are unrestricted by vocabulary or grammar and finds the best matches, even if some words deviate from the title. When searching EPG programs users may not know the full program name, or the recognizer or audio environment may loose a word. This allows us to take advantage of SpokenQuery's strength in combining probabilities over all the terms of a query and over all of the EPG entries to provide the user the best matches. To help the user understand this behavior we have combined SpokenQuery with result highlighting according to relevance.
Future Direction: We plan to refine the prototype and conduct further user testing. Advanced voice browsing techniques will also be investigated.
Contacts:
Derek Schwenke
Anthony Vetro
| Technical Reports: | |
| The Prospects for Unrestricted Speech Input for TV Content Search | |
| The MERL SpokenQuery Information Retrieval System - A System for Retrieving Pertinent Documents from a Spoken Query | |
Technology Areas:
Multimedia
Advanced Digital Television
Spoken Language Interfaces
Modification Date: July 3, 2007
