3D Face Recognition Based on Multi-Features of Face Image

カテゴリ: 部門大会

論文No: GS6-4

グループ名: 【C】平成22年電気学会電子・情報・システム部門大会講演論文集

発行日: 2010/09/02

タイトル(英語): 3D Face Recognition Based on Multi-Features of Face Image

著者名: Cuicui Zhang(Kumamoto University),Keiichi Uchimura(Kumamoto University),Go Koutaki(Kumamoto University),I Gede Pasek Suta Wijaya (Kumamoto University)

著者名(英語): Cuicui Zhang(Kumamoto University),Keiichi Uchimura(Kumamoto University),Go Koutaki(Kumamoto University),I Gede Pasek Suta Wijaya (Kumamoto University)

キーワード: 3D face recognition|mesh simplification|maximal principal curvature image|average edge image|range image

要約(日本語): An algorithm named 3D face recognition based on multi-features of face image is proposed in this paper. Recognition is performed based on three kinds of feature facial images: maximal principal curvature image, average edge image, and range image. The comparisons of the effect of mesh simplification to recognition rate are also shown in this paper. The experiments on a 3D facial database consisting of nearly 400 samples, which belongs to our lab show that our proposed method achieved excellent results. The highest accuracy rate of recognition based on range image achieved about 99% with just a small ratio of the original data.

PDFファイルサイズ: 6,416 Kバイト