No |
171405 |
標題(和) |
Facial Expression Analysis by Supervised ICA with a Selective Prior |
標題(英) |
Facial Expression Analysis by Supervised ICA with a Selective Prior |
研究会名(和) |
画像工学, 通信方式 |
研究会名(英) |
Image Engineering, Communication Systems |
開催年月日 |
2005-12-12 |
終了年月日 |
2005-12-13 |
会議種別コード |
5 |
共催団体名(和) |
|
資料番号 |
CS2005-57, IE2005-112 |
抄録(和) |
Feature selection is required when using ICA in feature extraction for pattern classification. Selection during ICA might provide a better candidate set of features. We propose a supervised ICA with a selective prior for the de-mixing coefficients so that those features with higher significance in discrimination could emerge easier during the learning. We formulate the learning rule for the supervised ICA in a form of the natural gradient approach and develop the algorithm of supervised ICA in facial expression analysis. The efficiency of the proposed algorithm has been investigated by numerical experiments. |
抄録(英) |
Feature selection is required when using ICA in feature extraction for pattern classification. Selection during ICA might provide a better candidate set of features. We propose a supervised ICA with a selective prior for the de-mixing coefficients so that those features with higher significance in discrimination could emerge easier during the learning. We formulate the learning rule for the supervised ICA in a form of the natural gradient approach and develop the algorithm of supervised ICA in facial expression analysis. The efficiency of the proposed algorithm has been investigated by numerical experiments. |
収録資料名(和) |
電子情報通信学会技術研究報告 |
収録資料の巻号 |
Vol.105, No.460 |
ページ開始 |
27 |
ページ終了 |
32 |
キーワード(和) |
|
キーワード(英) |
Facial Expression Recognition,ICA,Feature Selection,Sparse Coding |
本文の言語 |
ENG |
著者(和) |
陳凡 |
著者(ヨミ) |
チン ファン |
著者(英) |
Fan Chen |
所属機関(和) |
北陸先端科学技術大学院大学 |
所属機関(英) |
Japan Advanced Institute of Sci. and Tech. |
著者(和) |
小谷一孔 |
著者(ヨミ) |
コタニ カズノリ |
著者(英) |
Kazunori Kotani |
所属機関(和) |
北陸先端科学技術大学院大学 |
所属機関(英) |
Japan Advanced Institute of Sci. and Tech. |