Research On Recognition Method Of Non-contact Finger Phalangeal Prints

With the accelerating development of economy and the rapid rhythm of society,personal identification occupies a very important position in all walks of life, and therequirement is more and more widely. However, traditional methods, for example,passwords、ID cards、work permits and so on, are not convenient to carry, easy to lose andexist other defects. So the biometrics identification technology like fingerprint、handgeometry characteristic of knuckle print、vein、iris which are based on human body arepopular among the masses due to its accuracy, stability,etc. However, a variety of userauthentication technologies have advantages and disadvantages in affinity, cost andperformance, so the exploration of biological feature recognition is still an important taskof new research in this field.Finger Phalangeal Prints feature as emerging biometric features, are marked on thefront surface of the finger, which are obvious and have relatively uniform. As with theother hand biometric hand, palm and vein, finger Phalangeal Prints are stable and not easyto mimic.The characteristics of finger Phalangeal Prints of different people have differentdistributions of the location information, and refer to the section lines with the other handcharacterized by easy integration characteristics can play the role of the primary screening,so finger Phalangeal Prints are widely used in different contexts biometric technologies.Therefore, the paper selects finger Phalangeal Prints as identification, and will makedetailed study of the key algorithm for feature extraction in the process of identification.The paper presents a detection method for the relationship between the relativepositions of finger phalangeal prints. For self-built gallery of thumb outwardly open、fourfingers closed hand images, the first step is to find out the fingertips and finger valleypoints by using the Step edge detection pretreatment methods. Then the whole image isrotated to a horizontal position in order to compare the relative positions of fingerphalangeal prints. Finally, we can find out the horizontal coordinate position of fingerphalangeal prints in four fingers except the thumb by using the Roof edge detection method. In order to verify the proposed method, the use of self-built small natural gallerydisplay test results showed that the extraction rate is up to98.035%, indicating that themethod is feasible and effective. This paper presents a method for the detection of therelative position relationship between finger phalangeal prints, the formation of aone-dimensional feature vector by using the method of finger phalangeal prints spatiallocation, matching identification, and then the algorithm is embedded in DSP, therecognition rate can reach about70%.