Clustering Optimal Bayes Algorithm In The Application Of Hand Vein Recognition Research

Hand vein recognition increasingly rapid development and become a new field of biology image recognition technology in recent years,much of our attention in various identification technologies.Better uniqueness and adaptability hand vein recognition is an important advantage,but there are corresponding disadvantages,for example,when sampling and identification procedures are still not high recognition rate and a low efficiency and other issues,thus solving accuracy and efficiency this and other issues will be the key technology into applications.Although Bayesian algorithm is widely used in the field of identification,one by one if the low recognition efficiency,the article will be discussed for several commonly used image segmentation algorithm,combined with Bayesian clustering algorithm optimized,to ensure the accuracy of certain under the premise of large amounts of data to improve the situation under hand vein recognition efficiency.After thorough discussion in contrast,the final selection of the niblack edge detection algorithm as the algorithm used,and then by the wavelet transform texture feature extraction,the uniform distribution of dimensionality reduction.Then we get the image texture features matrix,and through the center of k clustering algorithm,the final recognition on Bayesian clustering results.The ultimate success of implementation of a large amount of data in the background,high recognition efficiency,accuracy,good recognition algorithm.The main work includes:(1)Trying to highlight the image texture information on a variety of image segmentation and edge detection algorithm is implemented and discuss,analyze advantages and disadvantages of various algorithms,including sobel,roberts,prewitt,log and other operators,as well as regional growth,Hough transform,threshold segmentation algorithm and niblack algorithm.And consideration from the visual effects,the effect of screening out a good and efficient algorithm.(2)Due to the large amount of data in the context of a single hand vein recognition efficiency is low,so the first set of images processed by the wavelet transform texture feature extraction,obtaining uniform distribution,etc.,for data dimensionality reduction,to give the result set matrix.Get a suitable number of clusters experience by way of trial and error will ultimately fall-dimensional data matrix is the most reasonable clustering.In order to achieve the purpose of reducing the recognition process sample space.(3)The minimum binary comparison algorithm and Bayes minimum risk Bayes algorithm advantages and disadvantages,and in theory discussions and proved the results.And then through an optimized algorithm for Bayesian hand vein texture recognition,more traditional identification methods and Bayesian clustering optimized Bayesian approach differences in the recognition rate,more traditional identification methods and Bayesian poly class optimized Bayesian approach in identifying efficiency differences and come to a final conclusion.