3D Palmprint Feature Extraction And Recognition Technology Research

Nowadays, information security has become a critic issue, and receives extensive attention. Physiological characteristic refers to data directly measured from the human body parts and some representative physiological features, including fingerprints, palmprint, hand shape, face, and so on. Palmprint as important biometric characteristic, has wide application prospects in human identity recognition because it is highly accurate and low cost. Most of the previous work have focused on two dimensional(2D) palmprint recognition in the past decade. This way loses the shape information when capturing palmprint images. Recently, 3D(Three-Dimension) biometrics have been widely studied and applied in human recognition. Unfortunately, the processing speed of 3D palmprint recognition systems is slow. The main reason is that information can’t obtain the real 3 D palmprint, feature extraction method of 3D data is not mature and so on.This paper presents a fast 3D palmprint capturing system and develops an efficient 3D palmprint feature extraction and recognition method. In order to fast acquire accurate 3D shape and texture of palmprint, a DLP(Digital Light Processing) projector triggers a CCD camera to realize synchronization. By generating a nd projecting green fringe pattern images onto the measured palm surface, 3D palmprint data are calculated from the fringe pattern images. The periodic feature vector can be derived from the calculated 3D palmprint data, so undistorted 3D biometrics is obtained. Firstly, the location and selection of ROI(Region O f Interest) are an important task in palmprint recognition. ROI will be used in the subsequent feature extraction and matching. The chosen ROI region will be 120x120 located in the palmprint center. Using the obtained 3D ROI and based on the definition of the mean curvature to get 3D palmprint every bit of curvature, 16 direction Gabor filter used to extract the image in the direction of the field, the direction of each point in the competition rules to determine the image field. Experimental results on capturing 3D palmprint show that 3D palmprint fast acquisition system can capture 12 images in a second and calculate for the shape of 3 D hand information. Recognition rate can reach 99.93%, matching in selecting the appropriate threshold. This method not only retains the 3D palmprint information to a great extent, improve the detection speed and recognition ability, but can effectively provide a reliable method of matching, 3D palmprint identification for the development of palmprint identification and technical route.