Voice: Workshop(s)
(Room Tampa Convention Center:Room 18-19)
20 Sep 16
4:30 PM
-
5:30 PM
Tracks:
: Technology & Enablers, : Technology & Enablers, : Technology & Enablers
Moderator(s):
Mark Przybocki
4:30-4:35 Group Organization
4:35-5:20 Group Discussion
5:20-5:30 Group Report Outs (overviews)
Group 1: Establishing a Large Scale (National) Database Creation
The group will consider issues associated with establishing a large-scale national database of known voice recordings for voice identity matching. To the greatest extent possible, its composition must support the establishment of realistic operating points for classifiers, and thus it must reflect realistic populations and collection conditions. The group may draw upon similar efforts and lessons learned through the process of establishing a national fingerprint database. Data broad enough to represent relevant populations, data capture conditions/channels, storage, ownership, recording content, quality, distribution/access, privacy, security, may all be topics the facilitator wishes to discuss.
Group 2: (research) Data Resources and Evaluation
This group will consider issues associated with the generation and distribution of the data resources necessary to reach a level of robustness in voice matching technology that meets the future state discussed in previous sessions. This group will consider strategic plans for establishing new data collections, integrity of ground truth, and prioritizing data collection needs.
Group 3: Promoting Technology Usage, Interoperability
This group will discuss accepted technology implementation protocols, standardization, best practices, score interpretation (meaningful operating points), and human operator accreditation