| . | ::::: |
Dataset: personalized web usage in developing countries
The first large scale, personalized web usage dataset in a developing country context. The data consists of enhanced browser logs collected over a period of 4 weeks between January and February of 2011 at an internet kiosk in the outskirts or Bangalore in Karnataka, India. It contains a total of 471 users, with 141 of those users having two or more sessions. For details, click here. |
| . | ::::: |
Source code: the hyke wrapper for biometric identification (voice recognition)
Hyke is a Python wrapper on top of two underlying C++ libraries that are used for several purposes in voice biometrics. Your first step should be to read the voice biometrics how-to, which is included in this distribution. This will give you a good idea on how the various pieces fit together in voice recognition. For details, click here. |
| . | ::::: |
Dataset: Telephone based speaker identification dataset from India
This audio data was collected for the purpose of speaker identification in developing country contexts. It includes a total of 83 unique voices, 35 female and 48 male collected over the telephone. In particular, it provides audio for performing limited vocabulary speaker identification using digit utterances. For details, click here. |