Palm print Recognition using PCA-based Adaptive Weighted Directional Features

Palm print recognition is a recent addition to the long list of biometric recognition which includes Iris, fingerprint, facial features and gait. The palm print unlike other biometric features is too many in numbers. This could lead to very long and exhaustive feature set. Also the minor deviation within the feature space tends to interfere with the accuracy of the palm print recognition. To overcome these existing drawbacks in this paper, we propose a novel PCA based adaptive weighting algorithm. The Principal Component Analysis (PCA) provides feature space compactness whereas the adaptive weighting suppresses the minor deviations that interfere with the performance. The proposed algorithm provides better accuracy than other existing state of the art methods. The Adaptive weighting is based on the orientation of edges. The proposed adaptive weighting PCA based palm print recognition (AWPCA-PR) is faster and efficient compared to other existing state of the art methods.