Research Of A Dynamic Handwriting Recognition Algorithm Based On DTW

With the rapid development of mobile Internet technology,smart phones,tablet PCs and other mobile devices have been popular.Mobile devices in the online shopping,online payment,transfer and other aspects of a convenient and fast features,online e-commerce trading volume is growing,leading to mobile platform e-commerce information security risks become more prominent,information security technology on the mobile platform Application is becoming more and more important.Identification based on biological characteristics has been widely used because it overcomes many defects of the traditional methods of identification.Handwriting is one of the important fields of biometric identification,it is accepted by the majority of users because of the advantages of simple equipment,easy collection,safety and the like.In this paper,the DTW algorithm and the efficient DTW algorithm are analyzed and improved,and a dynamic handwriting recognition algorithm based on DTW is proposed to solve the problem of handwriting handwriting on mobile platform.The main achievements of the thesis are as follows:1)This paper focuses on the basic research of dynamic handwriting recognition technology,carries on the theoretical and experimental analysis to the classical DTW algorithm and the efficient DTW algorithm,sums up the advantages and disadvantages of the two algorithms,which is beneficial to the improvement of the algorithm.2)The matching path of the DTW algorithm tends to the horizontal axis or the vertical axis,which leads to the large computational complexity of the algorithm,which has a great effect on the efficiency of handwriting recognition.This paper eliminates the path search method of the horizontal or vertical axis in the original algorithm,and proposes a cross-point search method,which reduces the computational complexity of the algorithm and effectively improves the efficiency of dynamic handwriting recognition.3)The effect of the preprocessing process on the recognition results is relatively large.By analyzing the shortcomings of overriding endpoint detection in the classical DTW algorithm,this paper proposes a method to relax the alignment of the starting end points and improve the performance of the algorithm,To improve the accuracy of dynamic handwriting recognition to provide a guarantee.4)In this paper,we combine the advantage of FastDTW algorithm to limit the slope of the path and change the slope of FastDTW algorithm.A new slope acquisition method is proposed to improve the recognition accuracy of dynamic handwriting recognition.5)In this paper,an improved algorithm based on DTW is proposed and verifiedexperimentally.The experimental results show that the error rate(FRR)and false positive rate(FAR)of the algorithm are 8.50% and 4.50% respectively,and the recognition efficiency is faster than that of the classical DTW algorithm.The recognition accuracy is higher than that of the FastDTW algorithm The rationality and feasibility of the algorithm in dynamic note recognition.