14-10-12 Lantian Li

From cslt Wiki
Revision as of 13:33, 13 October 2014 by Lilt (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Weekly Summary

1. Comparison of Two Methods for Kmeans Clustering

1)Means_model: Building virtual speaker models by calculating means.

2) Real_model: Searching for the nearest model around each virtual model.

Results show that the two methods is similar.

2. Design three-dimensional score vectors, each dimension of the vector is the system-score, tnorm-score, crank.

3. By means of SVM, making a binary classification of all score vectors.

Next Week

1. Statistical classification results, comparion with the eer result.

2. Expecting to use MLP to make classification .