An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition
In this paper, we investigate a feature conditioning method for the VTS-based model compensation. The VTS is a technique that predicts noisy acoustic model from clean acoustic model and noise model. It is noted that most of the previous studies use a single Gaussian noise model, which is unable to m...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2013
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/97488 http://hdl.handle.net/10220/11868 |
| _version_ | 1826129278812553216 |
|---|---|
| author | Li, Haizhou Nguyen, Duc Hoang Ha Xiao, Xiong Chng, Eng Siong |
| author2 | School of Computer Engineering |
| author_facet | School of Computer Engineering Li, Haizhou Nguyen, Duc Hoang Ha Xiao, Xiong Chng, Eng Siong |
| author_sort | Li, Haizhou |
| collection | NTU |
| description | In this paper, we investigate a feature conditioning method for the VTS-based model compensation. The VTS is a technique that predicts noisy acoustic model from clean acoustic model and noise model. It is noted that most of the previous studies use a single Gaussian noise model, which is unable to model noise statistics well, especially in non-stationary noisy environments. In this paper, we propose a combination of feature processing and VTS model compensation to handle non-stationary noise more efficiently. In the feature processing stage, the non-stationary characteristics of noise is reduced, hence the processed features is more suitable for VTS model compensation using single Gaussian noise model. Experimental analysis on the AURORA2 task shows that the proposed method has the potential to improve the performance of VTS method in non-stationary environments if good noise estimation is available. |
| first_indexed | 2024-10-01T07:38:01Z |
| format | Conference Paper |
| id | ntu-10356/97488 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:38:01Z |
| publishDate | 2013 |
| record_format | dspace |
| spelling | ntu-10356/974882020-05-28T07:17:16Z An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition Li, Haizhou Nguyen, Duc Hoang Ha Xiao, Xiong Chng, Eng Siong School of Computer Engineering International Symposium on Chinese Spoken Language Processing (8th : 2012 : Kowloon, Hong Kong) Temasek Laboratories DRNTU::Engineering::Computer science and engineering In this paper, we investigate a feature conditioning method for the VTS-based model compensation. The VTS is a technique that predicts noisy acoustic model from clean acoustic model and noise model. It is noted that most of the previous studies use a single Gaussian noise model, which is unable to model noise statistics well, especially in non-stationary noisy environments. In this paper, we propose a combination of feature processing and VTS model compensation to handle non-stationary noise more efficiently. In the feature processing stage, the non-stationary characteristics of noise is reduced, hence the processed features is more suitable for VTS model compensation using single Gaussian noise model. Experimental analysis on the AURORA2 task shows that the proposed method has the potential to improve the performance of VTS method in non-stationary environments if good noise estimation is available. 2013-07-18T05:59:35Z 2019-12-06T19:43:14Z 2013-07-18T05:59:35Z 2019-12-06T19:43:14Z 2012 2012 Conference Paper Nguyen, D. H. H., Xiao, X., Chng, E. S., & Li, H. (2012). An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition. 2012 8th International Symposium on Chinese Spoken Language Processing (ISCSLP). https://hdl.handle.net/10356/97488 http://hdl.handle.net/10220/11868 10.1109/ISCSLP.2012.6423503 en © 2012 IEEE. |
| spellingShingle | DRNTU::Engineering::Computer science and engineering Li, Haizhou Nguyen, Duc Hoang Ha Xiao, Xiong Chng, Eng Siong An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition |
| title | An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition |
| title_full | An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition |
| title_fullStr | An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition |
| title_full_unstemmed | An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition |
| title_short | An analysis of vector Taylor series model compensation for non-stationary noise in speech recognition |
| title_sort | analysis of vector taylor series model compensation for non stationary noise in speech recognition |
| topic | DRNTU::Engineering::Computer science and engineering |
| url | https://hdl.handle.net/10356/97488 http://hdl.handle.net/10220/11868 |
| work_keys_str_mv | AT lihaizhou ananalysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition AT nguyenduchoangha ananalysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition AT xiaoxiong ananalysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition AT chngengsiong ananalysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition AT lihaizhou analysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition AT nguyenduchoangha analysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition AT xiaoxiong analysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition AT chngengsiong analysisofvectortaylorseriesmodelcompensationfornonstationarynoiseinspeechrecognition |