Image Acquisition And Local Wave Processing Method Research On Hand Vein

Recognition of face and fingerprint are widely used. But face recognitionsystem is expensive, broken finger skin and conditions of different moistureaffect the range of application of fingerprint recognition. Hand vein recognitionbased on vein characteristics in the skin under the back of hand can overcomethe above drawbacks, and prevents the forgery of identification featureseffectively. One of the key steps in hand vein recognition is to extract the handdorsal vein from hand vein image accurately. Different from ordinary imaging,hand vein image is scattered light imaging, it is blur and has no clear vascularstructure. In this paper, a new extraction method which is much more accurateand clearer is proposed based on the analysis of existing method of hand dorsalvein extraction.The main work is summarized as follows:(1) According to the particularity and requirement of acquisition of handvein images, the near infrared LED light source which can be adjusted flexibly isdesigned to improve the acquisition quality as far as possible.(2) Aiming at the difference of the gray image caused by differentplacement height and sloping angle in the process of acquisition, a series ofpretreatment are taken, the inconsistency of position and sloping angle of thehand in the process of acquisition are solved. These measures include extractionof effective area and normalization of size and gray. Then the effective area afternormalized is enhanced by using different methods.(3) The existing extraction methods of hand dorsal vein are summarized: theextraction method to test hand dorsal vein is based on edge detection. Threshold segmentations method is based on template. And the extraction method is basedon wavelet analysis. In this paper, these existing methods are tested. Due to theresidual image from Local Wave decomposition can represent the basiccharacteristics of the original image. So the method based on the characteristicwith Fourier Transform is proposed in order to improve the extraction accuracyof hand dorsal vein. Compared with the traditional extraction method in detail,the new method is more accurate.(4) Aiming at the problems of information redundancy of hand dorsal veinfrom Local Wave decomposition with Fourier Transform, a series of measuresfor post processing are adopted in the paper, then the interference of inessentialcomponents is eliminated, and the vein information is much better.The drawbacks of traditional vascular extraction algorithm is overcame byusing the method of Local Wave decomposition with Fourier Transform, as aresult, main outline of the vein is highlighted, unnecessary feature information ofvascular edge is cut more accurately, the information of structure that need to beidentified is enhanced and the display of detail characteristics is also clearer.