| Palm-print recognition rises in recent years. and soon becomes the focus of biometric identification. Palm-print image contains very rich information, it is hand to wear, this is better than fingerprints:Palm-print recognition has higher accuracy than the face recognition; Signature recognition will be subject to their own emotional and external factors, so palm-print recognition has higher stability. However, palm-print recognition does not have the widely development than iris recognition, face recognition and fingerprint recognition as its late development.The design of a palm-print recognition system involves image data collection, image preprocessing, feature extraction, feature matching. This paper concentrates on the deep research of the image preprocessing and feature extraction in palm-print identification system.The positioning of palm-print is a very important part in palm-print image processing, its results will directly affect the following feature extraction and feature matching. This paper analyzes the palm positioning segmentation method used, due to the particularity of palm-print images, these methods are not suitable in practical applications. In this paper, a new algorithm based on the classical algorithm considers the contour information and local area information, and become more suitable in palm-print recognition. Finally, experimental results demonstrate the effectiveness of the algorithm.Based on the wavelet multi-resolution properties, the energy of the palm lines and papillary pattern distribute at different scales. Palm lines do not have the oscillation, the wavelet decomposition will increase as the scale. Papillary pattern has some oscillation, large-scale wavelet coefficients is much smaller than the small-scale. According to this, palm wavelet energy is extracted as the feature in this paper. Meanwhile, the proportion of each component energy will be a good feature. At last, experimental results show the effectiveness of this method. |
PDF Full Text Request