Research On Digital Watermarking Encryption Algorithm For Electronic Bill Authentication

Digital image watermarking uses image processing,cryptography,and information hiding to achieve image authentication and protection.It has important application value in the field of financial security.Through the digital watermark encryption algorithm,it can effectively improve the performance of digital watermarking,and the requirement for the authentication of the ticket image in the actual financial system can be satisfied,thereby ensuring the security and the efficiency of encryption and decryption.This paper first analyzes the research status and several challenges in the field of digital watermark encryption.Combined with the specific needs and technical bottlenecks in the process of financial authentication,this paper deeply studies the digital watermark encryption algorithm for financial electronic bill authentication system from the following three aspects.Several key technologies for watermark embedding and authentication:1)digital watermark embedding and extraction in binary images;2)anti-compression robust electronic bill watermarking algorithm;3)detection and multi-encryption of key areas of financial instruments.The main work and innovative achievements of the thesis include:1.Aiming at the fax of electronic bill in financial system,the watermark embedding algorithm in binary image is studied.This paper analyzes the edge information extraction of binary image,studies the method of embedding watermark with edge information,and combines chaotic mapping to scramble the watermark,and studies the effect of different chaotic mapping on watermark detection.The encryption and decryption of the watermark embedded in the binary image are experimentally verified.During the embedding process,chaotic mapping is used to encrypt the watermark further,which effectively improves the security.2.This paper analyzes the attack method of digital watermark,deeply analyzes the JPEG image compression standard,and proposes a method to resist JPEG compression attack according to the relationship between wavelet subband coefficients.Based on the relative relation of wavelet subband coefficients,robust watermark embedding is performed by chaotic encryption using the constructed wavelet coefficient matrix and binary sequence.From the angle of image peak to signal-to-noise ratio and image compression rate,the watermark's attack ability is improved.At different image compression rates,the embedded watermarking pattern is little affected,which effectively improves the anti-attack of electronic bills in transmission.Experimental results show that the proposed algorithm can effectively resist JPEG compression attacks and improve the security of electronic bill authentication.Experiments show that the chaotic sequence encrypted digital watermarking algorithm has better robustness against JPEG compression.3.Aiming at the special needs of financial bills,this paper proposes a method to detect the key information area of bills and multi-watermark encryption,and proposes a key area encryption algorithm based on neural network and image feature matching.This algorithm combines visible and invisible multi-watermark to authenticate and protect electronic bills,which improves tampering and forgery.The strength of detection.Visible watermarking can effectively protect the source information of bills,directly protect bills from tampering,and invisible watermarking can protect sensitive information locally.The combination of visible and invisible watermarking can be more effective in anti-counterfeiting authentication of bills.It can detect and encrypt the key area of the bill,and has strong ability of tamper-proof,and effectively ensures the security of the financial bill.4.Combining with the depth learning theory and the ability of convolution operation to extract features from images,a method of extracting watermark using depth convolution network is proposed.Experiments show that this method can significantly reduce the error rate compared with the traditional algorithm in the case of low SNR or received attacks.This paper proposes to use the bill graph and its differential features to construct multi-channel image input features suitable for CNN networks,and fully exploit the intrinsic features of the image;further improve the traditional CNN network structure,and connect the output of all convolution layers into one layer,including each The fusion feature of layer information is input into the fully connected layer of the network for classification and identification.Experimental results show that the improved CNN recognition algorithm,compared with the traditional CNN,DNN and other algorithms,can more efficiently identify the content of multiple key areas,so as to effectively encrypt multiple digital watermarks,improve the security and operational efficiency of financial paper processing.