Eigenspectra Palmprint Recognition

This paper introduces a novel technique for palmprint recognition on the transform domain, based on combining principle component analysis (PCA) and Fourier domain. Principal Component Analysis (PCA) has been widely adopted as the most promising biometric recognition algorithm. However, has its limitations: poor discriminatory power in the presence of variant illumination and large computational load when the original dimensionality is high and the number of training sample is large. Traditionally, to represent the palmprint image, PCA is performed on the whole spatial image. where in the proposed method, Fourier transform is used to decompose an image into its sine and cosine component, then the spectrum is used for PCA representation since doing PCA on the whole frequency domain does not achieve any performance. In comparison with the traditional use of PCA and three other methods, the proposed method gives better recognition accuracy and discriminatory power using one training imag...

• Type: Conference
• Year: 2008
• Where: DELTA
• Authors: Moussadek Laadjel, Ahmed Bouridane, Fatih Kurugollu
• Keywords: Component Analysis, DELTA 2008, Discriminatory Power, Fourier Transform, Information Management