| 電子情報通信学会 研究会発表申込システム 講演論文 詳細 |
技報閲覧サービス [ログイン] 技報アーカイブ |
| トップに戻る | 前のページに戻る | [Japanese] / [English] |
| 講演抄録／キーワード | ||
| 講演名 | 2012-02-09 11:30 Face recognition based on virtual frontal view generation using LVTM with local patches clustering ○Xi Li（Nagoya Univ.）・Tomokazu Takahashi（Gifu Shotoku Gakuen Univ.）・Daisuke Deguchi・Ichiro Ide・Hiroshi Murase（Nagoya Univ.）  PRMU2011-191 SP2011-106 |
|
| 抄録 | （和） | One of the major difficulties encountered by face recognition is the varying poses caused by in-depth rotations. The intra-person appearance differences caused by rotations are often larger than the inter-person differences, which makes the traditional face recognition methods such as eigen-face infeasible. This paper presents a framework for face recognition across pose based on virtual frontal view generation using Local View Transition Model(LVTM) with local patches clustering. Previous study on LVTM shows that more accurate appearance transition model can be achieved by first dividing the original face image plane into overlapping local patch regions and then the learned transition models for each patch are aggregated for the final transformation. In this paper we show that the accuracy the appearance transition model and the recognition rate can be further improved by better exploiting the inherent linear relationship between frontal-nonfrontal face image pairs. This is achieved based on the observation that variations in appearance caused by pose are closely related to the corresponding 3D face structure and intuitively frontal-nonfrontal pairs from more similar local 3D face structures should have a stronger linear relationship. For each specific location, instead of learning a common transformation as in LVTM, the corresponding local patches are first clustered based on appearance similarity distance metric and then the transition models are learned separately for each cluster. In the testing stage, each local patch for the input nonfrontal probe image is transformed using the learned local view transition model corresponding to the most visually similar cluster. The experimental results on real life face dataset demonstrate the effectiveness of the proposed method. |
| （英） | One of the major difficulties encountered by face recognition is the varying poses caused by in-depth rotations. The intra-person appearance differences caused by rotations are often larger than the inter-person differences, which makes the traditional face recognition methods such as eigen-face infeasible. This paper presents a framework for face recognition across pose based on virtual frontal view generation using Local View Transition Model(LVTM) with local patches clustering. Previous study on LVTM shows that more accurate appearance transition model can be achieved by first dividing the original face image plane into overlapping local patch regions and then the learned transition models for each patch are aggregated for the final transformation. In this paper we show that the accuracy the appearance transition model and the recognition rate can be further improved by better exploiting the inherent linear relationship between frontal-nonfrontal face image pairs. This is achieved based on the observation that variations in appearance caused by pose are closely related to the corresponding 3D face structure and intuitively frontal-nonfrontal pairs from more similar local 3D face structures should have a stronger linear relationship. For each specific location, instead of learning a common transformation as in LVTM, the corresponding local patches are first clustered based on appearance similarity distance metric and then the transition models are learned separately for each cluster. In the testing stage, each local patch for the input nonfrontal probe image is transformed using the learned local view transition model corresponding to the most visually similar cluster. The experimental results on real life face dataset demonstrate the effectiveness of the proposed method. |
|
| キーワード | （和） | face recognition / cross pose / view transition model / local patch / clustering / / / |
| （英） | face recognition / cross pose / view transition model / local patch / clustering / / / | |
| 文献情報 | 信学技報, vol. 111, no. 430, PRMU2011-191, pp. 31-36, 2012年2月. | |
| 資料番号 | PRMU2011-191 | |
| 発行日 | 2012-02-02 (PRMU, SP) | |
| ISSN | Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 | |
| 著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します．(許諾番号：10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) | |
| PDFダウンロード | PRMU2011-191 SP2011-106 |
|
| 研究会情報 | |
| 研究会 | PRMU SP |
| 開催期間 | 2012-02-09 - 2012-02-10 |
| 開催地（和） | 東北大学 青葉山キャンパス |
| 開催地（英） | |
| テーマ（和） | 時系列パターン認識 |
| テーマ（英） | |
| 講演論文情報の詳細 | |
| 申込み研究会 | PRMU |
| 会議コード | 2012-02-PRMU-SP |
| 本文の言語 | 英語 |
| タイトル（和） | |
| サブタイトル（和） | |
| タイトル（英） | Face recognition based on virtual frontal view generation using LVTM with local patches clustering |
| サブタイトル（英） | |
| キーワード(1)（和/英） | face recognition / face recognition |
| キーワード(2)（和/英） | cross pose / cross pose |
| キーワード(3)（和/英） | view transition model / view transition model |
| キーワード(4)（和/英） | local patch / local patch |
| キーワード(5)（和/英） | clustering / clustering |
| キーワード(6)（和/英） | / |
| キーワード(7)（和/英） | / |
| キーワード(8)（和/英） | / |
| 第1著者 氏名（和/英/ヨミ） | 李 析 / Xi Li / シ リ |
| 第1著者 所属（和/英） | 名古屋大学 (略称： 名大) Nagoya University (略称： Nagoya Univ.) |
| 第2著者 氏名（和/英/ヨミ） | 高橋 友和 / Tomokazu Takahashi / タカハシ トモカズ |
| 第2著者 所属（和/英） | 岐阜聖徳学園大学 (略称： 岐阜聖徳学園大) Gifu Shotoku Gakuen University (略称： Gifu Shotoku Gakuen Univ.) |
| 第3著者 氏名（和/英/ヨミ） | 出口 大輔 / Daisuke Deguchi / デグチ ダイスケ |
| 第3著者 所属（和/英） | 名古屋大学 (略称： 名大) Nagoya University (略称： Nagoya Univ.) |
| 第4著者 氏名（和/英/ヨミ） | 井手 一郎 / Ichiro Ide / イデ イチロウ |
| 第4著者 所属（和/英） | 名古屋大学 (略称： 名大) Nagoya University (略称： Nagoya Univ.) |
| 第5著者 氏名（和/英/ヨミ） | 村瀬 洋 / Hiroshi Murase / ムラセ ヒロシ |
| 第5著者 所属（和/英） | 名古屋大学 (略称： 名大) Nagoya University (略称： Nagoya Univ.) |
| 講演者 | 第1著者 |
| 発表日時 | 2012-02-09 11:30:00 |
| 発表時間 | 30分 |
| 申込先研究会 | PRMU |
| 資料番号 | PRMU2011-191, SP2011-106 |
| 巻番号（vol） | vol.111 |
| 号番号（no） | no.430(PRMU), no.431(SP) |
| ページ範囲 | pp.31-36 |
| ページ数 | 6 |
| 発行日 | 2012-02-02 (PRMU, SP) |
[研究会発表申込システムのトップページに戻る]