The Information Hiding Technology And Its Application For Encrypting Printed Face Image And Fingerprint

With new kinds of digital image processing software appeared, and the print,scanning, photographic equipment are to be upgraded every years, the certificate andtrademark counterfeit, the fake prints piracy are more easier. There are growing demandfor copyright identification of an image.The combination of optical technology and thedigital watermarking technique is one way of solving this problem.The copyright identification technology of two special kinds of image like the faceimage and fingerprints was discussed in this paper. Four kinds of watermarking schemehave been explored. The first is embeding invisible anti-counterfeiting mark byprojecting encryption beam instead of traditional digital watermarking technique intocolor portrait photo.The second is print-scan resilient watermarking technology.Thethird is multiple watermarks embeddeing technology including embeddeing multiplewatermarks into portrait and embeddeing multiple watermarks into fingerprint image.The research contents include:1. Embeding invisible anti-counterfeiting mark based on Fourier transform byprojecting encryption beam into color portrait photo.This paper gives a new method of adding watermark into portraiture photography(printed). The combination of liquid crystal display(LCD) projectors and camera,instead of traditional computer image encryption procedures, was used for embedding a"simulation watermark" to the color photos. The characteristic is embedding watermarkdirectly at the moment of exposure.The experimental results show that by using analog/digital mixing technology,security identification markers can be embedded in Fourier domain of the color image.The watermark is resistant after printing, scanning and being local damaged.2. Transforming fingerprint image into encryption graphics basing on double phaseencryption and discrete fractional Fourier transform(DFRT) principle is used for thecertificate security.The first method: After scrambling binary fingerprint image through Arnoldtransform. and additional fractional Fourier transform, the gray encryption graphicsconsidering fingerprint information can be got. Experimental results show that thefingerprint mark still can be extracted from1/4area damage or dirty.The second method: weighted-type fractional Fourier transform amplitudespectrums of two images including fingerprint image and personal information imageare added. Experimental results show that the fingerprint mark and personal informationimage still can be extracted from1/4area damage or dirty (as a result of two of theimage mutual interference, the watermark extraction was mixed with the background).3. Embeddeing multiple watermarks in an image is presented. (a) Embeddeing multiple fingerprint watermarks into portrait based on fractionalFourier transform principle.A new method of multiple watermarks embedding simultaneously based ondiscrete fractional Fourier transform is presented. The different watermarks areconverted one-dimension sequences and embedded simultaneously on differentpositions of one-dimension fractional Fourier transform middle part amplitude spectrumbase on different orders. We get the corresponding watermark image through two keys:the orders and the positions. When one of them is not correct, the watermark image can’tbe obtained. And the technology needn’t the original image when we extractivewatermark images, so it is an blind watermark images embedding technology. Whenseveral people have the rights of the image work, the different fractional Fouriertransform orders and the different positions are distributed to different people. And italso can be used for embedding different characteristic watermark images in an image.(b) Embeddeing multiple watermarks including portrait image into fingerprintimage based on wavelet transform and fractional Fourier transform principle.The fingerprint image watermark embedding technology has two great difficulties:(1) The original fingerprint image is simple, it is perceive easily after embedding thewatermark;(2) The geometric distortion caused by the watermark fingerprint can lead toits original identity authentication function failure.We solve preliminarily the two technical difficulties based on wavelet transform,fractional Fourier transform principle and the least significant bit(LSB) technology, andmake the embedded watermark smaller interference.4. Embeddeing multiple watermarks simultaneously for identifying image statusbased on wavelet transform, singular value decomposition(SVD) technology、LSBtechnology and DFRT principle is presented. According to the extracting situation ofwatermarks from an image and the twice DFRT amplitude spectrum, we can identify theimage situation including copyright ownership, and whether is tampered, printed orscanned.