“140915 - Lantian Li”版本间的差异

2014年9月16日 (二) 02:14的版本

Weekly Summary

1. To help Thomas write the paper "an overview of robust speaker recognition" and finish the first part

"Applcations"and the second part "speaker-dependent variablilty", and then integrate the fisrt version.

2. To go on studying a scoring method on GMM-UBM aiming to design a cohort reference speaker models.

Rearrange a test set and clustering training set. Complete data preparation task .

Design the K-means algorithem to cluster the training set in order for organizing the cohort set, but still have bugs.

Next Week

1.Go on the task2 including K-means designing and scoring distribution statistics.