The NIST Open Speech-Activity-Detection evaluation (OpenSAD) was intended to provide Speech-Activity-Detection (SAD) system developers with an independent evaluation of performance on a variety of audio data. The purpose of a SAD system is to find regions of speech in an audio file. The purpose of the evaluation was to advance SAD technology and broaden the research community.
Documentation
Changes in version 10 of the Evaluation Plan
- Updated to reflect what actually happened in the OpenSAD evaluation
Scoring software
Schedule (dates are all in 2015)
- September 3: Publish registration form, mock data, and beta-version scoring code
- September 7: Release data-license agreements, from LDC
- Sept. 21: Release Training and DevTest datasets and release beta-version scoring code
- Oct. 1: Participant registration deadline
- Nov. 2-13: Evaluation period (two weeks to accommodate multi-site teams)
- Nov. 26: Release preliminary results (automated, non-refereed)
- Dec. 18: One-day OpenSAD workshop at NIST (Gaithersburg, MD)
- End of December: Release official results
Contact
opensad_poc [at] nist.gov (opensad_poc[at]nist[dot]gov)