http://index.cslt.org/mediawiki/index.php?action=history&feed=atom&title=Lantian_Li_2015-04-08Lantian Li 2015-04-08 - 版本历史2026-04-09T04:57:45Z本wiki的该页面的版本历史MediaWiki 1.23.3http://index.cslt.org/mediawiki/index.php?title=Lantian_Li_2015-04-08&diff=14588&oldid=prevLilt：以“Weekly Summary 1. Addtional experiments are made for semi text-independent SV based on d-vector. Results show that phone posteriors (without utterance CMVN but mak...”为内容创建页面2015-04-08T14:25:08Z<p>以“Weekly Summary 1. Addtional experiments are made for semi text-independent SV based on d-vector. Results show that phone posteriors (without utterance CMVN but mak...”为内容创建页面</p>
<p><b>新页面</b></p><div>Weekly Summary<br />
<br />
1. Addtional experiments are made for semi text-independent SV based on d-vector.<br />
<br />
Results show that phone posteriors (without utterance CMVN but make Global CMVN) <br />
<br />
combined with Fbanks (without utterance CMVN but make Global CMVN) as DNN input feature<br />
<br />
performs the best. <br />
<br />
2. Addtional experiments are made for text-dependent SV on d-vector. Aiming to compare the performance between <br />
<br />
CNN and DNN. <br />
<br />
Results show that the performance of d-vectors extracted from the traditional DNN are better than CNN's on this text-dependent SV task.<br />
<br />
3. Prepare for the cohort-based classfication task in score-level on GMM-UBM system. <br />
<br />
Next Week<br />
<br />
1. Go on the task 2, 3.</div>Lilt