Presentation 2013-12-21
Mode Estimation by Mean Shift Learning with Variable Window Width and Gaussian Assumption and The Application to Background Sound Spectrogram Estimation
Tomokazu TAGUCHI, Yasunari YOKOTA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) There are many techniques for mode estimation, such as graph spectrum and mean shift. Generally, these conventional techniques have some fundamental problems caused of determination of kernel window width and convergence to local optima. Although some methods have tried to overcome these problems, it was resulted in increasing calculation cost. Introducing assumption that main cluster follows Gaussian distribution loosely, a mean shift technique in which kernel window width can be estimated iteratively and is utilized for mode estimation. This article investigated consistency and unbiasedness of the technique numerically, and applied to background sound spectrogram estimation.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Mode estimation / kernel window width determination / graph spectrogram / mean shift / background sound spectrogram
Paper # NC2013-55
Date of Issue

Conference Information
Committee NC
Conference Date 2013/12/14(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 Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Mode Estimation by Mean Shift Learning with Variable Window Width and Gaussian Assumption and The Application to Background Sound Spectrogram Estimation
Sub Title (in English)
Keyword(1) Mode estimation
Keyword(2) kernel window width determination
Keyword(3) graph spectrogram
Keyword(4) mean shift
Keyword(5) background sound spectrogram
1st Author's Name Tomokazu TAGUCHI
1st Author's Affiliation Faculty of Engineering, Gifu University()
2nd Author's Name Yasunari YOKOTA
2nd Author's Affiliation Faculty of Engineering, Gifu University
Date 2013-12-21
Paper # NC2013-55
Volume (vol) vol.113
Number (no) 374
Page pp.pp.-
#Pages 6
Date of Issue