Face recognition across pose using self-organizing maps (ニューロコンピューティング)

スポンサーリンク

概要

• 論文の詳細を見る

• We present a comparison between Self-organizing map (SOM) and hierarchical SOM (HSOM) to tackle pose variation problem in face recognition. Self-organizing map (SOM) transforms the high dimensional face images into low dimensional topological space. However, in SOM, the hierarchical relation between data set does not preserve. HSOM solve this problem, moreover it reduces the time required for winner searching. In this paper, CMU-PIE face database is utilized to show the accuracy of both models.

• 社団法人電子情報通信学会の論文
• 2007-03-09

著者

• Taniguchi Rin-ichiro

  Department Of Intelligent Systems Kyushu University

• Taniguchi Rin-ichiro

  Department Of Advanced Information Technology Kyushu University

• Tsuruta Naoyuki

  Department Of Electronics Engineering And Computer Science Fukuoka University

• ALY Saleh

  Department of Intelligent Systems, Kyushu University

• Aly Saleh

  Department Of Intelligent Systems Kyushu University

関連論文

• Hybrid background modeling for long-term and short-term illumination changes (特集 ビジョン技術によるイノベーション)
• A Randomized Self-Organizing Maps for Gesture Recognition (論文特集 自己組織化マップとその応用)
• 変形股関節を対象とした多階層統計的形状モデルによる3次元CT画像からの大腿骨領域の半自動抽出
• On using the Self Organizing Map in Face Recognition
• Fast Learning Approach Using Self Organizing Map(一般セッション(1),文字認識・文書理解)
• Fast Learning Approach Using Self Organizing Map(一般セッション(1), 文字認識・文書理解)
• Shape and Pose Parameter Acquisition of 3D Multi-Part Objects on Multiple Viewpoint Image Sequences
• Face recognition across pose using self-organizing maps (ニューロコンピューティング)

スポンサーリンク