Palmprint Recognition Via Localized DCT Feature Extraction

カテゴリ: 部門大会

論文No: MC5-3

グループ名: 【C】平成21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: 2009/09/03

タイトル(英語): Palmprint Recognition Via Localized DCT Feature Extraction

著者名: Hillary Choge(徳島大学),尾山 匡浩(徳島大学),Stephen Karungaru(徳島大学),柘植 覚(徳島大学),福見 稔(徳島大学)

著者名(英語): Hillary Choge(University Of Tokushima),Tadahiro Oyama(University Of Tokushima),Stephen Karungaru(University Of Tokushima),Satoru Tsuge(University Of Tokushima),Minoru Fukumi(University Of Tokushima)

キーワード: palmprint recognition|DCT|localized feature extraction

要約(日本語): The use of biometrics for personal identification has recently attracted enormous attention from researchers, and the palmprint has steadily gained prominence for this purpose. However, most research has so far focused mainly on global structural or texture feature extraction for palmprint recognition. In this paper, we present a novel method that exploits the well known capability of the Discrete Cosine Transform to represent natural images using only a few coefficients to efficiently extract local palmprint features, and concatenate them into a compact feature vector that represents each palmprint. Recognition and verification experiments using these feature vectors show that this method can effectively be used to discriminate between different palmprints and offers a viable alternative to the available techniques.

PDFファイルサイズ: 9,397 Kバイト