Photo
Caption Me Now

Caption Me Now

 

Project Summary:
The Caption Me Now technology addresses the need to get a transcription of audio or video materials, by focusing on the rapid and cost-effective provision of transcription services.

IBM Caption me now can be used in the captioning of lectures and Webcasts to make them accessible to deaf users. The service uses IBM's Super Human speech recognition technology and the ViaScribe captioning interface to make captioning services more pervasive without incurring large costs. Existing solutions are costly, relying on manual transcription of audio content by highly skilled stenographers. IBM Caption me now automates a large percentage of the transcription with the remaining content being transcribed by stenographers.

The prototype of the technology was demonstrated at the 2004 annual IBM Stockholders meeting with great success, and became an IBM First-Of-A-Kind (FOAK) prototype in January of 2004.

Objectives:
Design, implement, and deploy an automatic and semi-automatic transcription framework, enabling on-demand captioning of audio materials.

Customer Problems:
Audio on the web is provided without an associated transcription & cannot be searched based on a user's particular interests; users are limited to the audio channel alone, & the information is unavailable for deaf/hard of hearing users that go to the site. This barrier violates the Americans with Disabilities Act (ADA) (amendment Section 508 of the Rehabilitation Act), that requires reasonable accommodations for people with disabilities.

Solution Description:
Web sites containing audio will be enhanced with a "Caption Me Now" button to activate a speech recognition system (ViaScribe) on a server that will create a transcript of the spoken words, which can also be re-integrated into the multimedia format as captioning.

Business Value:
Provides significant service revenues for IBM and differentiates IBM's capabilities Can spawn real-time transcription & automatic summarization of meetings, machine translation and search of audio archives.

Publication List