API descriptionΒΆ
Copyright 2014-2020 Anthony Larcher and Sylvain Meignier
- Authors
Anthony LACHER, Sylvain MEIGNIER & Kong Aik LEE
- Version
1.3.1 of 2019/01/22
This package is the core of the SIDEKIT toolkit.
While developing SIDEKIT, we tried to keep in mind two
targets:
1. limit the number of classes to allow better readability
2. make SIDEKIT compatible with existing tools
To reach this target, we have created four Main Classes
(
FeaturesExtractor`
, FeaturesServer
, Mixture
and StatServer
) which can be usedtogether with a number of tools available in companion
modules (
sidekit_io
and sv_utils
).Front-end and back-end processing such as acoustic feature extraction
The
frontend
package include a number of tools to extract and normalizethe acoustic features as well as detecting the high energy frames for voice
activity detection. The
bosaris
package consists of the translation ofa part of the BOSARIS toolkit available on this webpage.
The current python implementation of the BOSARIS toolkit does not include tools
for calibration and fusion but only the core structures that are used to manage
enrollment lists, trial definitions and scores.
The authors would like to thank Niko Brummer and AGNITIO to allow them to distribute
this version of the BOSARIS toolkit.
Note
The bosaris package which is released together with SIDEKIT is distributed under a different license. The intellectual property belongs to the original authors of the toolkit.