An exploration of robust model structures for image variations based on statistical approaches

• Project/Area Number: 19700165
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Single-year Grants
• Research Field: Perception information processing/Intelligent robotics
• Research Institution: Nagoya Institute of Technology
• Principal Investigator: NANKAKU Yoshihiko Nagoya Institute of Technology, 大学院・工学研究科, 助教 (80397497)
• Project Period (FY): 2007 – 2009
• Project Status: Completed (Fiscal Year 2009)
• Budget Amount: ¥3,710,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥510,000); Fiscal Year 2009: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000); Fiscal Year 2008: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000); Fiscal Year 2007: ¥1,500,000 (Direct Cost: ¥1,500,000)
• Keywords: パターン認識 / 画像認識 / 統計モデル / 隠れマルコフモデル / 特徴抽出 / 確率的主成分分析 / 因子分析

Research Abstract

The objective of this research is to find general model structures which are robust to image variations for image recognition. Separable lattice HMMs (SL-HMM) have been proposed as statistical models which can deal with location and size variations. This research proposed a new model structure which integrates linear feature extraction based on probabilistic principal component analysis and factor analysis into SL-HMMs. Furthermore, SL-HMMs were extended to models which can represent rotation variations and which include explicit state duration models.

Research Products

• [Journal Article] Face recognition based on separable lattice 2-D HMM with state duration modeling (2010)
  • Author(s): Yoshiaki Takahashi, Akira Tamamori, Yoshihiko Nankaku, Keiichi Tokuda
  • Journal Title: 2010 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2010) Pages: 2162-2165
  • NAID: 110007327156
  • Peer Reviewed
• [Journal Article] An extension of separable lattice 2-D HMMs for rotational data variations (2010)
  • Author(s): Akira Tamamori, Yoshihiko Nankaku, Keiichi Tokuda
  • Journal Title: 2010 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2010) Pages: 2206-2209
  • NAID: 10031126716
  • Peer Reviewed
• [Journal Article] 状態継続長を考慮した分離型2次元格子HMMによる顏画像認識 (2009)
  • Author(s): 高橋良彰, 玉森聡, 南角吉彦, 徳田恵一
  • Journal Title: 電子情報通信学会技術研究報告 vol.108,no.484 Pages: 153-158
• [Journal Article] 回転変動を考慮した分離型2次元HMMによる顔画像認識 (2009)
  • Author(s): 玉森聡, 南角吉彦, 徳田恵一
  • Journal Title: 電子情報通信学会技術研究報告 vol.108,no.484 Pages: 159-164
• [Journal Article] Face recognition using hidden Markov eigenface models (2007)
  • Author(s): Yoshihiko Nankaku, Keiichi Tokuda
  • Journal Title: 2007 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2007) Vol.2 Pages: 469-472
  • Peer Reviewed
• [Presentation] An extension of separable lattice 2-D HMMs for rotational dato variations (2010)
  • Author(s): Akira Tamamori, Yoshihikc Nankaku, Keiichi Tokuda
  • Organizer: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)
  • Place of Presentation: ダラス(アメリカ)
  • Year and Date: 2010-03-19
• [Presentation] Face recognition based on separable lattice 2-D HMM with state duration modeling (2010)
  • Author(s): Yoshiaki Takahashi, Akira Tamamori, Yoshihiko Nankaku, Keiichi Tokuda
  • Organizer: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)
  • Place of Presentation: ダラス(アメリカ)
  • Year and Date: 2010-03-18
• [Presentation] 状態継続長を考慮した分離型2次元格子HMMによる顔画像認識 (2009)
  • Author(s): 高橋良彰, 玉森聡, 南角吉彦, 徳田恵一
  • Organizer: 電子情報通信学会技術研究報告
  • Place of Presentation: 東北工業大学
  • Year and Date: 2009-03-14
• [Presentation] 回転変動を考慮した分離型2次元HMMによる顔画像認識 (2009)
  • Author(s): 玉森聡, 南角吉彦, 徳田恵一
  • Organizer: 電子情報通信学会技術研究報告
  • Place of Presentation: 東北工業大学
  • Year and Date: 2009-03-14
• [Presentation] クロスバリデーションを用いたベイズ基準によるコンテキストクラスタリング (2008)
  • Author(s): 橋本佳, 全炳河, 南角吉彦, 李 晃伸, 徳田恵一
  • Organizer: 日本音響学会2008年春季研究発表会
  • Place of Presentation: 千葉
  • Year and Date: 2008-03-18
• [Presentation] 時系列マッチングを含む統計モデルに基づいた継続長およびスペクトルの同時変換 (2008)
  • Author(s): 油谷かおり, 宇藤陽介, 南角吉彦, 戸田智基, 李 晃伸, 徳田恵一
  • Organizer: 日本音響学会2008年春季研究発表会
  • Place of Presentation: 千葉
  • Year and Date: 2008-03-18
• [Presentation] Model-Space MLLR for Trajectory HMMs (2007)
  • Author(s): Heiga Zen, Yoshihiko Nankaku, Keiichi Tokuda
  • Organizer: Interspeech 2007-EUROSPEECH
  • Place of Presentation: アントワープ(ベルギー)
  • Year and Date: 2007-08-30
• [Presentation] Spectral conversion based on statistical models including time-frequency matching (2007)
  • Author(s): Yoshihiko Nankaku, Kenichi Nakamura, and Keiichi Tokuda
  • Organizer: ISCA Speech Synthesis Workshop
  • Place of Presentation: ボン(ドイツ)
  • Year and Date: 2007-08-24
• [Presentation] Face recognition using hidden Markov eigenface models (2007)
  • Author(s): Yoshihiko Nankaku, and Keiichi Tokuda
  • Organizer: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)
  • Place of Presentation: ハワイ(アメリカ)
  • Year and Date: 2007-04-20