Automatic Identification Algorithm Of Texture Anti-Counterfeiting Tag Base On Transform Domain

In recent years, counterfeiting caused serious harm to consumers. It has been a serious social and political issue. While anti-counterfeitingtechnology has been an effective way to anti-counterfeit.Texture anti-counterfeiting is one kind of new anti-counterfeiting technology developed in recent years, it’s a collection of database storage, digital printing,information queries and other technologies in one. It’s highly personalized comprehensive way to anti-counterfeit. In addition, compared with the traditional security theory, the texture of texture anti-counterfeitingtag is randomly generated by the paper. Let counterfeiters difficult to counterfeit and maintain long-term effective. At the same time it combined with the networkand communication technology and easy for consumers to identify the query. Texture anti-counterfeitingtechnology has become the emergence and development of new anti-counterfeiting technology in recent years.However, due to the current market texture anti-counterfeiting technology is hard to distinguishin poor light and other external environmental conditions by eyes. Therefore, the automatic way to identify is imperative.In this paper, the main achievement of this paper is automatic identification algorithm of texture anti-counterfeiting tag base on transform domain, the main work are as follows:(1) Wepropose that automatic identification algorithm of texture anti-counterfeiting tag base on transform domain,the proposed algorithm has strong robustness. It deal with anti-counterfeiting texturelabels by two-dimensionaltransformation, such as DCT、 DFT and DWT-DFT.Then, the visual feature vectors can be extracted from transform domain by two-dimensionaltransformation. At the same time, we create a database of texture images’ feature vectors. In the end of the scheme,we can ensure the authenticity of a productby comparing the value of Normalized Cross-correlation between the feature vectorsfrom the database. The results of experimental study demonstrate that the proposed algorithm has strong robustness against common and geometric attacks.And italso has strong robustness through shoot by different models of mobile phones.(2)As wejust store the feature vectors in the database, so we can greatly reduce the database storage space occupied.This paper compared the Normalized Cross-correlation (NC) between the feature vectors to identify the texture anti-counterfeiting automatically and quickly.