Palmprint Recognition Method Based On Rotation-Invariant

With the rapid development of information, biometric feature is a novel method in personal identity verification, which is substituted for the traditional identification technology. Due to the palmprint has the lifelong invariance and uniqueness, attention of a large number of scholars. At present, the theory of palmprint recognition methods are improved gradually.Palmprint recognition technologies include: the acquisition process of palmprint image、pretreatment process、palmprint feature extraction and feature recognition. Feature extraction is one of most basic problems in the research of palmprint recognition. Extracting effective palmprint feature is the crucial problem in the field of palmprint recognition. Palmprint image has rich information which includes structural characteristics of line and obvious texture features. There were research focus on how to select the feature, the lacking of main orientation for palmprint recognition system led to an incorrect feature extraction and feature matching. In order to extract more precise palmprint feature, a new method for feature extraction of palmprint is proposed in the research of palmprint recognition-based on rotation-invariant method of palmprint feature extraction.(1) For the problem of the inaccurate positioning of the key point, which caused by the error of extracting the resign of interest,a new preprocessing algorithm based on fixed point is presented in this paper. This method first cuts the corner area of the palm image, then finds a baseline and counts the distance between all the pixels and the baseline. Last, cutting out the region of interest.(2) The paper designs the palmprint recognition method. The method extracts palmprint feature by using the histogram of oriented gradient firstly, but the algorithm ignores the direction information of the image, so the method of histogram of oriented gradient and dominant orientation has been combined in this paper, and obtained directional feature of each sub-block.(3) The paper introduces the palmprint recognition method based on rotation-invariant, whose further improvements in robustness has been obtained. The method improves the recognition results and enable the palmprint identification system has rotation invariant.(4) Before the process of palmprint recognition, encoding extracted features which using a binary scheme according to the thought of coding, and obtaining two groups of binary-coded images, which is simple to implement during the identification.(5) Through MATLAB GUI interface to make each stage of processing visualization, and build a practical platform. This system has following functions: image reading、image storage、image processing、feature extraction and feature matching.