Project : metiss
Audio segmentation and classification toolkit
In the framework of our activities on audio indexing and speaker recognition, audioseg, a toolkit for the segmentation of audio streams, was developed this year. This toolkit provides generic tools for the segmentation and indexing of audio streams under Unix, such as audio activity detection, abrupt change detection, segment clustering, Gaussian mixture modeling and joint segmentation and detection using hidden Markov models. The toolkit relies on SPro, a free software developed by a member of our team, for feature extraction and should be itself distributed as a free software.
The audioseg toolkit has been used to develop a new speaker verification platform, validated with our participation to the NIST speaker recognition evaluation this year . It was also extensively used for our work on the detection of audio events in video sound tracks  .
Contact : firstname.lastname@example.org