| Presentation | 2008-05-29 Acceleration of Higher-Order ICA Using Lerning Period Detection Based on Closed-Form Second-Order ICA and Kurtosis Yuuki FUJIHARA, Yu TAKAHASHI, Shigeki MIYABE, Hiroshi SARUWATARI, Kiyohiro SHIKANO, Akira TANAKA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | A hands-free system with conventional independent component analysis(ICA) should update the separation filter constantly to follow the hourly environment change. In an actual environment, however, there exists a time period when the user is absent and only noises arise. If we continuously conduct the separation-filter updating in such a period via the same manner when target speech exists, ICA often yields an improper separation filter and performance degradation of the hands-free system. This study proposes a novel algorithm of kurtosis-based voice activity detection (VAD) for an appropriate ICA optimization. In the proposed algorithm, first, closed-form 2nd-order ICA (SO-ICA) is performed for providing the roughly-separated signals, and based on the kurtosis of the roughly-separated signal, VAD is processed. Next, when the current time period is voiced block, higher-order ICA's re-optimization is applied using the previously obtained SO-ICA's solution as the initial filter; this results in a fast and high convergence. In this paper, we reveal the effectiveness of the proposed method by performing a simulation experiment for blind spatial subtraction array with the proposed method under heavily noisy condition. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Independent component analysis / blind source separation / voice activity detection / closed-form solution |
| Paper # | EA2008-15,SIP2008-24 |
| Date of Issue |
| Conference Information | |
| Committee | EA |
|---|---|
| Conference Date | 2008/5/22(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 | Engineering Acoustics (EA) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Acceleration of Higher-Order ICA Using Lerning Period Detection Based on Closed-Form Second-Order ICA and Kurtosis |
| Sub Title (in English) | |
| Keyword(1) | Independent component analysis |
| Keyword(2) | blind source separation |
| Keyword(3) | voice activity detection |
| Keyword(4) | closed-form solution |
| 1st Author's Name | Yuuki FUJIHARA |
| 1st Author's Affiliation | Nara Institute of Science and Technology() |
| 2nd Author's Name | Yu TAKAHASHI |
| 2nd Author's Affiliation | Nara Institute of Science and Technology |
| 3rd Author's Name | Shigeki MIYABE |
| 3rd Author's Affiliation | Nara Institute of Science and Technology |
| 4th Author's Name | Hiroshi SARUWATARI |
| 4th Author's Affiliation | Nara Institute of Science and Technology |
| 5th Author's Name | Kiyohiro SHIKANO |
| 5th Author's Affiliation | Nara Institute of Science and Technology |
| 6th Author's Name | Akira TANAKA |
| 6th Author's Affiliation | Graduate School of Information Science and Technology, Hokkaido University |
| Date | 2008-05-29 |
| Paper # | EA2008-15,SIP2008-24 |
| Volume (vol) | vol.108 |
| Number (no) | 68 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |