| Presentation | 2012-05-24 Development of robust voice activity detection using empirical mode decomposition and modulation spectrum analysis Yasuaki KANAI, Masashi UNOKI, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Voice activity detection (VAD) is used to detect speech/non-speech periods in observed signals. However, current VAD technique has a serious problem in that the accuracy of detection of speech periods drastically reduces if it is used for noisy speech and/or for mixtures of speech/non-speech. We proposed a robust VAD using empirical mode decomposition (EMD) and modulation spectrum analysis (MSA) to resolve these problems (method A). EMD is used to decompose the observed signals into intrinsic mode functions (IMFs). Then, speech/non-speech periods are determined from resynthesize non-stationary IMFs by using MSA. Method A detected speech periods more correctly than conventional method. However, False detection was not able to be lost completely. This paper examined the techniques that raise robustness of method A. We proposed VAD that is detected from each IMFs by MSA(method B). As a result of evaluation, method B was able to do VAD which was more exact and stabilize (detection accuracy is not influenced by an analysis signal) than the method A. It has checked that robustness was improving. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | voice activity detection / empirical mode decomposition / modulation spectrum analysis / false acceptance rate / false rejection rate |
| Paper # | EA2012-1,SIP2012-1,SP2012-1 |
| Date of Issue |
| Conference Information | |
| Committee | SIP |
|---|---|
| Conference Date | 2012/5/17(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Signal Processing (SIP) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Development of robust voice activity detection using empirical mode decomposition and modulation spectrum analysis |
| Sub Title (in English) | |
| Keyword(1) | voice activity detection |
| Keyword(2) | empirical mode decomposition |
| Keyword(3) | modulation spectrum analysis |
| Keyword(4) | false acceptance rate |
| Keyword(5) | false rejection rate |
| 1st Author's Name | Yasuaki KANAI |
| 1st Author's Affiliation | School of Information Science, Japan Advanced Institute of Science and Technology() |
| 2nd Author's Name | Masashi UNOKI |
| 2nd Author's Affiliation | School of Information Science, Japan Advanced Institute of Science and Technology |
| Date | 2012-05-24 |
| Paper # | EA2012-1,SIP2012-1,SP2012-1 |
| Volume (vol) | vol.112 |
| Number (no) | 48 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |