Font Size: a A A

Research Of The Digital Image Atermarking Algorithm Based On Region Of Interest

Posted on:2014-11-19Degree:MasterType:Thesis
Country:ChinaCandidate:Q E MaFull Text:PDF
GTID:2268330401973012Subject:The computer system structure.
Abstract/Summary:PDF Full Text Request
With the development of Internet, digital multimedia works include image, audio, videoet al. are published and spread on the internet with the infringement problem. The copyrightprotection of digital media has become a hot research topic at home and abroad. An effectivetechnical method to solve the copyright of digital media problem is the digital watermarkingtechnique. Through several years’ development, watermark technique in regions of interest(ROI) becomes an important research direction. Based on image understanding, it findsimportant part of the image, and then embeds the watermark. Compared to traditionalwatermark which was embedded on entire carrier, it is off course a significant improvement.The thesis mainly studies watermark technique based on color image’s ROI, the mainwork include:(1) comprehensively introduce the traditional watermarking techniques andimage’s ROI, and how to combine them for application;(2) detailed study the application ofwavelet transform in image watermarking;(3) extract the color image’s ROI, and according tothe different characteristics of the image’s three channels, select corresponding algorithms toembed the watermark in the image of channel B and channel G at the same time.First it is a general introduction of watermark technique from its definition, classification,characteristics, technique of watermark generating, embedding and detecting, watermarkquality evaluation, watermark attack and the countermeasure; Then, it is a brief introductionof ROI technique from its concept and classification, and then emphatically introduce Itti’ssaliency map model which is based on visual features.Second with the study of the wavelet transform domain watermark, we get thewatermark embedding strategy, that is, with its enough watermark capacity, it should select ashigh as possible wavelet decomposition level, and the selection of watermark embedding areais the low frequency region. Then, a suitable quantization watermarking scheme in wavelettransform domain is proposed, which is in order to achieve blind watermark detection, andachieve self-adaptability of the algorithm based on the tree structure and wavelet blocks inwavelet transform, and so improve the watermark result.Last color image and ROI based watermarking system embeds the watermark on theimage of channel B in its time domain and channel G in its wavelet transform domain at thesame time; through their cooperation, it is robust to various attacks for the whole watermarking system. From the result of simulation experiment by Matlab, it finds that thesystem has good robustness for attack of cropping, noise, low pass filtering, rotating, scaling,JPEG and JPEG2000compression et al.In this study, the image watermarking technique combines the image ROI technique andwavelet transform technique, which makes the watermarking system more perfect and morepractical.
Keywords/Search Tags:image watermarking, wavelet transform, regions of interest, robustness
PDF Full Text Request
Related items