Analysis And Research On Detection Algorithm Of Digital Image Area Copy And Paste Tamper Detection

With the development of information technology, image information has become an important processing object in the communication and computer systems, however, with the appearance of various powerful image processing software, people can tamper and forge data information of image arbitrarily by the software, and if these false image is used illegally, such as judicial identification, military intelligence, news media, bank notes and other false image information, it will result in serious impact and losses to individuals, society and the state. Therefore, this paper carries out analysis and research on digital image tamper detection algorithm.For the same regional digital image which is copy-paste tampered, adopting the image tamper detection and location algorithm based on the texture and moment invariant feature. On the basis of basic image processing algorithm, using texture characteristic and moment invariant characteristic combined with wavelet algorithm to realize the sorting, similarity matching and mathematical morphology description of the characteristics, applying threshold discrimination to achieve detection and location. The experiments show that, the proposed scheme can not only precisely detect and locate the copy-paste tampering region of the same image, but also resist the noise pollution, JPEG compression damage and translation, scaling, rotation and so on attacks. Moreover, the algorithm has low complexity, high detection efficiency and good robustness.For the digital image which is copy-paste tampered, adopting the improved multi-scale invariant feature of image tampering detection and location algorithm. Using multi-scale auto-convolution invariant feature, SIFT feature and SURF feature combined with wavelet algorithm to realize the feature extraction and analysis, and to sort, match similarity and describe the mathematical morphology of the characteristics. Finally, use threshold discrimination to achieve detection and location. The experiments show that, the proposed improved scheme can not only precisely detect and locate the copy-paste tampering region of the image, but also resist the Gauss noise, JPEG compression and translation, scaling, rotation and so on attacks. Also, it can solve the problem of information loss of the feature point, and the algorithm has low complexity, high detection efficiency and good robustness.For the digital image which is copy-paste tampered, adopting image tamper detection and recovery algorithm based on the wavelet transform and phase correlation method. Realize the image restoration technique based on wavelet transform, first, using the method of regional searching, realize image matching by the way of phase correlation, then, fusing the image information. The experiments show that, the proposed scheme can not only precisely detect, locate and recover the copy-paste tampering region of the image, but also resist post processing operations, such as fuzzy, noise, fuzzy plus noise attack. It can reduce the computational cost and improve the efficiency of image restoration.