Research On Identity Recognition Based On Dynamic Lip Features

With the development of the internet and the application of big data,the importance of information security has become increasingly prominent.Traditional identity recognition technology cannot meet the increasing requirements for its security.Therefore,how to improve the security of identity recognition has become a hot topic of research.At this background,identity recognition technology based on biological characteristics has received widespread attention due to its relatively high security.Biological characteristics include human face,fingerprint,gait,etc.These characteristics are unique to each person.Therefore,the identity recognition method based on biological characteristics is not easy to lose,can be used anytime and anywhere,and is difficult to be forged,which improves the security of identity recognition.As a kind of biological characteristic,the lip features has unique physiological characteristics,and the dynamic lip features also reflects acquired behavior characteristics,contains more information,and can better reflect individual differences.Therefore,this paper will realize the identity recognition based on dynamic lip features.This paper proposes an identity recognition method based on dynamic lip features,focusing on solving three problems,including: lip positioning,dynamic lip features extraction,and dynamic lip features recognition.Among them,for the lip positioning,this paper uses a method based on face detection combined with the geometric features of facial features to determine the lip area;for the extraction of dynamic lip features,this paper uses the method of image difference to obtain the lip motion dynamic difference image,and use morphological processing and edge detection and extraction methods to process the contour features of the dynamic lip;for the recognition of dynamic lip features,this paper gives four lip features recognition methods,which are based on cosine similarity,based on HOG features,based on LBP features,and based on Hu invariant moments,and conducted a comparative analysis.And the results show that the lip features recognition method based on Hu invariant moments is more effective,and the accuracy rate is above 85%.Finally,this paper conducts overall experiments on the proposed identity recognition method based on dynamic lip features.The results show that the method can effectively realize the identification of the person to be tested,and has a higher success rate and a lower false recognition rate.