Note: Your custom schedule will not be saved unless you create a new account or login to an existing account.
| Paper: | SP-P13.11 |
| Session: | General Topics in Robust Speech Recognition |
| Time: | Thursday, May 20, 13:00 - 15:00 |
| Presentation: |
Poster |
| Topic: |
Speech Processing: Robust Speech Recognition |
| Title: |
FUSION BASED SPEECH SEGMENTATION IN DARPA SPINE2 TASK |
| Authors: |
Chengyi Zheng; OGI School of Science & Engineering | | |
| | Yonghong Yan; OGI School of Science & Engineering | | |
| Abstract: |
We report a new fusion based segmentation approach using multiplefilter bank coefficients. This new approach takes advantage ofcurrent feature extraction procedure, with little additionalcomputation cost. Another level of fusion was performed bycombining several segmentation systems. Evaluation was conductedon the second SPeech In Noisy Environments (SPINE2) task.Experiments show our fusion based approaches significantly reducedthe WER compare to two classifier-based approaches. Compare to themanual segmentation, our approach only has 0.3\% WER increase. |
| |
| Back | |