- From: Jim Larson <jim@larson-tech.com>
- Date: Thu, 29 Jan 2004 09:11:10 -0800
- To: judith@jmarkowitz.com
- Cc: www-voice@w3.org, scott.McGlashan@hp.com, 'Cathy Tilton' <ctilton@saflink.com>
Judith, Thanks for your suggestions for linking a new version of VoiceXML to BioAPI, the standard for biometrics. I will bring this suggestion to the Voice Browser Working Group as they define the requirements for the the the follow on to VoiceXML 2.0 Regards, -Jim jmarkowitz wrote: >Dear Committee Members > >I understand from Jim Larsen that you plan to begin work on VoiceXML >Version 3 in the near future. I also understand that you want to >incorporate new features and functions into that versions of VoiceXML. > >I strongly urge the committee to select speaker authentication as one of >those new functions. This brief proposal explains why and suggests a >method for incorporating speaker authentication that is fast yet brings >to bear some of the best thought on building secure, well-designed >applications. > >Background >Interest in and use of speaker authentication has been growing rapidly. >Private industry is adopting the technology for things like password >reset and government is interested in using it to facilitate telephone >access to sensitive information and for Homeland Security projects. This >year will be a banner year for government work because funding for >Homeland Security projects will begin to flow to contractors. Biometrics >-- including voice -are important components of a number of those >projects. > >The biometric industry has responded by creating a number of standards. >The standard that is of greatest relevance and utility for VoiceXML is >BioAPI. BioAPI is a high-level, generic API for biometrics. Among the >objectives of creating BioAPI were to enhance interoperability of >products and systems and to promote multi-biometric (multi-modal) >deployments. Multi-biometric applications take many forms, including >those that support the use of different biometrics for different facets >of a larger security system. For example, a bank might use face or >fingerprint recognition to access account information at an ATM but >employ speaker authentication to access the same information by >telephone. > >BioAPI is an international standard. It was originally developed by a >consortium organized by IBM, Compaq, Novell, and leading biometric >vendors. One of the earliest implementations included >speaker-verification products. I have attached a copy of BioAPI 1.1 for >your consideration. > >In 2002, the BioAPI specification was approved (with some modifications) >by ANSI under the name INCITS 358-2002. ISO, of which ANSI is a member, >"internationalized" it and, by the end of 2003, the ISO working >committee had completed its final draft (ISO/IEC Final Committee Draft >19784). > >Proposal >I urge the committee to review the attached document and to move to >incorporate speaker authentication into VoiceXML Version 3 by creating a >link to BioAPI. Linking VoiceXML to BioAPI would bring a number of >important benefits to the committee, the speech-processing industry, and >VoiceXML. They include: > >. Dramatically shorten and simplify the standardization process for > incorporating speaker authentication into VoiceXML - allowing > the committee to focus on other new features >. Ensure that issues related to building voice-biometric systems have > been addressed by individuals who understand both biometrics and > security >. Facilitate the design of standards-based, multi-biometric deployments >. Satisfy standards requirements in RFPs for biometrics >. Bring the speech-processing industry into alignment with the rest of >the > biometrics industry. > >Since I learned about the plan to work on VoiceXML Version 3 last fall, >I've asked several speaker-verification vendors and integrators to >review BioAPI 1.1 to determine whether there would be anything in it >that would be problematic for their products and technologies. They have >all reported they found nothing of concern in the specification. > >I would be happy to answer any questions that the committee might have >and to put the committee into contact with Cathy Tilton, of the Chair of >BioAPI Consortium. Ms. Tilton has expressed support for the idea of >linking the specifications. > >Thank you for your consideration of this proposal. > >Sincerely, > >Judith Markowitz, Ph.D., president >J. Markowitz, Consultants >5801 North Sheridan Road, Suite 19A >Chicago, IL 60660 >Tel: 773-769-9243 >Fax: 773-769-9253 >Email: Judith@JMarkowitz.com > > > > >
Received on Thursday, 29 January 2004 12:12:05 UTC