The Research Of3D Shape And Texture Acquisition Technology Of Palmprint And Hand

Palmprint and hand shape as two kinds of important biometric characteristics, have wideapplication prospects in human identity recognition because of their uniqueness, life-longinvariance, not easily lost, good anti-fake performance and other characteristics.The existing research about palmprint and hand shape just use part of the hand features, andare mostly based on2D images. However, palmprint and hand shape are3D biometriccharacteristics. When using the acquired2D image as the carrier of biometrics, the nonlinearmapping from3D to2D inevitably loses information or causes distortion, so that the subsequentfeature extraction and feature matching are based on inaccurate or incomplete input information.Moreover, palmprint and hand shape are in the same position of the human body, so using thetwo bometric characteristics together will greatly increase the accuracy.Therefore, this paper presents a method to acquire accurate3D shape and color texture of thepalmprint and hand shape simultaneously. The composite RGB fringe pattern image is projectedon the surface of a hand through DLP projector. From a different viewpoint, a CCD cameracaptures the deformed fringe patterns with respect to the hand shape. After compensating for thecrosstalk and chromatic aberration between color channels, three fringe patterns are extractedfrom the red, green and blue channels of the captured composite color images. Calculated theabsolute phase of the hand by using four-step phase shifting method and optimum three-fringenumber selection method, so absolute phase map is obtained. Finally, the color textureinformation can be directly captured or demodulated from the fringe pattern images. The systemis calibrated build up the relationship between absolute phase map and3D shape. Therefore,accurate3D shape and color texture of the palmprint and hand shape were acquiredsimultaneously. Experimental results show that the proposed acquisition method can get3Dshape and color texture information of the palmprint and hand shape, In addition, the largestabsolute error in X, Y, Z direction is not more than0.0535mm. Therefore, the acquired highaccurate3D shape and texture information of the palmprint and hand shape will improvestability, accuracy and integrity of the identity identification system.