Lantian Li 15-06-16
From cslt Wiki
Revision as of 04:16, 16 June 2015 by Lilt
1. Database label on digital Speech synthesis and Replay.
2. Accomplish the codes of Speech sequences' correlation coefficient.
3. Prepare for three deep speaker embedding tasks:
1). large-scale deep speaker vector framework: Complete the phone decoding process. 2). derive binary i-vectors using Hamming distance learning： results as shown in CVSS 373. It seems that block-training binary representation is interesting.
1. Go on the task 3.