Robust Palmprint Recognition Base On Contactless Palmprint Acquired

Today's society is a highly information and the rise of the network society, so the security of information and network is very important. Traditional methods of biometric authentication are easy to be lost, counterfeited and imitated, greatly reducing the security of the system and information. Biometric technology can effectively overcome the disadvantages of the methods of traditional identification. Palmprint recognition is a biometric technology which has a user-friendliness and convenient application environment. Thus, in recent years, it has been widely developed.Under non-contact acquisition mode, because of the user's hand is not touching any platform, palmprint images may not be distorted for pressure and do not have public sanitary issue. In addition, users do not worry latent palm prints which remain on the sensor's surface could be copied for illegitimate uses. However, contactless palmprint recognition system brings an unstable environment for imaging, position of hand is mutative that increases difficulty of locating Region of Interest (ROI) of palmprint. Especially, variant illumination have seriously affected on the ability of the system to recognize individuals. In order to improve the robustness of contactless palmprint system, expand the application of palmprint recognition, touch-less palmprint recognition system can be used in varying illumination and complex background conditions. A robust palmprint recognition approach base on contactless palmprint acquired is proposed. Firstly, the hand image is segmented from the background by using the skin-color thresholding method. Then, a novel valley detection algorithm is used to find the valleys of the fingers. These valleys serve as the base points to locate the ROI of the palm. Next, the local binary patterns (LBP) which is robustness to illumination variation is applied to extract the texture feature of the palmprint. Finally, feature matching is performed by using chi square statistic and modified PNN (probabilistic neural network). Experiments show that this approach can expand the application of palmprint recognition; the promising results can be achieved in varying illumination and complex background base on contactless palmprint acquired.In order to further study the robustness of contactless palmprint recognition system, a set of simulated of contactless on-line palmprint recognition system using of functions of Image Acquisition Toolbox and GUI are provided by Matlab software is designed in this paper. Through palmprint image on-line acquisition, location, feature extraction and matching process, the purpose of identification can be achieved.