| Presentation | 2006-03-22 Omni-directional Estimation of Sound Source Location and Its Application to Estimation of Speaker's Position by Combining with Skin-color Information Satoshi TAKAHASHI, Jun-ichi IMAI, Masahide KANEKO, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | A user doesn't always stand in front of a robot and may sometimes call a robot from its back. Therefore a robot should know the user's position first to start communication with him/her. This paper proposes the omni-directional estimation method of speaker's position using the combination of audio and color information. Estimation of the position of sound source is carried out to calculate the difference of arrival time from the sound source to multi-channel microphones. Number of microphones and their optimal arrangement are derived considering the accuracy and processing time for 3-D estimation of sound source position. Next the robust detection of skin-color region is carried out by combining a trained GMM (Gaussian Mixture Model) for input scene with a general GMM. Bayesian network is employed to combine the result of sound source estimation and detection of skin-color region, and to realize a highly accurate estimation of speaker's location. Experimental results are shown to demonstrate the usefulness of the proposed methods. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | estimation of sound source location / omni-direction / speaker's location / skin-color information / CSP method / Gaussian mixture model |
| Paper # | MVE2005-70 |
| Date of Issue |
| Conference Information | |
| Committee | MVE |
|---|---|
| Conference Date | 2006/3/15(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 | Media Experience and Virtual Environment (MVE) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Omni-directional Estimation of Sound Source Location and Its Application to Estimation of Speaker's Position by Combining with Skin-color Information |
| Sub Title (in English) | |
| Keyword(1) | estimation of sound source location |
| Keyword(2) | omni-direction |
| Keyword(3) | speaker's location |
| Keyword(4) | skin-color information |
| Keyword(5) | CSP method |
| Keyword(6) | Gaussian mixture model |
| 1st Author's Name | Satoshi TAKAHASHI |
| 1st Author's Affiliation | Graduate School of Electro-Communications, The University of Eletro-Communications() |
| 2nd Author's Name | Jun-ichi IMAI |
| 2nd Author's Affiliation | Graduate School of Electro-Communications, The University of Eletro-Communications |
| 3rd Author's Name | Masahide KANEKO |
| 3rd Author's Affiliation | Graduate School of Electro-Communications, The University of Eletro-Communications |
| Date | 2006-03-22 |
| Paper # | MVE2005-70 |
| Volume (vol) | vol.105 |
| Number (no) | 683 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |