Study On Palmprint Image Positioning Segmentation And Hierarchical Main Line Feature Extraction

Today, advances in information technology, while providing people with a variety ofconvenient, but also there are a variety of security risks. Compared to the traditionalsecurity technology, and biometric identification technology through science andtechnology with biometric combination of human inherent physiological characteristicsor behavioral characteristics identity The confirmation of the reliability of the securitytechnology.Palmprint identification system as an emerging biometric technology, has attractedincreasing attention because of their unique characteristics. Contains a wealth of featureinformation in the palm, there are both suitable for high-resolution image, minutiaefeatures for line features in the low-resolution images of these characteristics databasefor the establishment of high standards, are very favorable conditions.This paper studies the ROI of the palm under the gray-scale image region extractionand palm main line extraction. The main research work include the following:1、The ROI region extraction of palmprintThe part of the first collection of images in the database based on the canny operatorbinarization denoising palm of the binary image; then using mathematical morphology,corrosion and expansion algorithm to extract the contours of the binary image, done toremove the empty and contour extraction process; after8-direction chain code contourimage, said by the characteristics of the analysis of contour chain code, find the featurepoints in the contour; given according to the last step feature points, the coordinatesystem needs to establish a crop the image ROI area, cut the palm of the ROI and ROIregion were normalized to get a128x128image.2、 the classification to extract the palm main lineIn many feature points of the palm, the palm of the three major flexors line that is thelifeline of wisdom lines and emotions, is the clearest stripes line is basicallyaccompanied by a person’s life does not change. Everyone is different, the stability andreliability are good. In fact, the palm of the most important feature is also a ridge characteristics, and in the low-resolution and low quality images is still able to clearlyidentify. And structural features of the ridge is almost free from the influence of light,noise and the palm of your hand placement, and have sufficient discriminant information.Due to the different aspects of the study of the palm main line, the method is alsochanging. Based on the relationship between the line and the energy, palm main line ofthe wavelet decomposition to extract the main line; through the design of nonlinearfilters to extract the palm; palm lines as a ridge line, according to a first and second Thefirst derivative to determine the palm.This paper presents a new extraction mainline. Before details of the new method, theimage gray, and two kinds of extraction methods based on the main line of gray. Oneapproach is to first palmprint sub-block, then the gray histogram to determine the localthreshold of each block, divided according to the threshold determined after The secondapproach is to consider the8neighborhood pixels and the pixel location, to determine thethreshold image segmentation. The analysis of these two methods is to determine thethreshold to extract the ridge, and in the same gray scale. By analyzing the different graylevel grayscale images, the paper presents the main line feature extraction method basedon gray-scale classification of the palm.