en-de
en-es
en-fr
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
Speech-to-Text Evaluation System
Published on Feb 25, 20073399 Views
Related categories
Chapter list
Further Progress in Meeting Recognition: The ICSI-SRI Spring 2005 Speech-to-Text Evaluation System 00:01
Overview 00:18
Data Sets00:46
Evaluation Tasks01:56
Development Strategy: Base System02:23
Development Strategy: Meeting System02:35
Acoustic Preprocessing02:53
Improved MDM Delay-Summing04:09
IHM Crosstalk Filtering05:08
“Fast” Decoding Architecture06:46
Full System Architecture07:56
Acoustic Features and Models09:19
Baseline Model Improvements10:54
Acoustic Adaptation11:37
MLP Feature Portability12:58
Combining CTS and BN Base Models13:31
Discriminative MAP Adaptation14:58
Test-time MLLR Improvement15:25
Conference Meeting LMs16:01
Lecture Meeting LMs18:14
LM Results (see paper for perplexity results)19:43
Overall Results: Conference Meetings21:38
Lecture System Development25:18
Lecture Recognition Results27:12
Conclusions28:51
Future Work29:52
Thank You!31:31