The main difference with traditional methods is that the proposed scheme first segments the test image in semantically independent patches before extracting key points. As a result, copy motion regions can be detected by matching these patches. The matching process consists of two stages. In the first stage, we find the suspect pairs of patches that may contain copy movement counterfeit zones, and we estimate approximately an affine transformation matrix. In the second step, an algorithm based on the Expectation Maximisation is designed to refine the estimated matrix and to confirm the existence of copy-move falsification. The experimental results show the good performance of the proposed scheme compared to the state-of-the-art schemes of the public databases.
The detection of image counterfeits is currently one of the fields of research interested in image processing. Counterfeiting copy and move (CM) is one of the most used techniques. In this paper, we propose a method that is efficient and fast to detect copy-move regions. The proposed method accelerates the block matching strategy. First, the image is divided into fixed size overlay blocks and then a discrete cosine transform is applied to each block to represent its characteristics. The k-means fast grouping technique is used to group the blocks into different classes. Zigzag scanning is performed to reduce the length of each vector of block characteristics. The characteristic vectors of each cluster block are lexicographically sorted by radix ordering, the correlation between each nearby block indicates its similarity. The experimental results show that the proposed method can efficiently detect duplicate regions and reduce the processing time to 50% of other previous works.
The availability of powerful digital image processing programs, such as PhotoShop, makes it relatively easy to create digital forgeries from one or several images. The White House was reshaped and blurred to create an illusion of an unfocused background. Then, Bill Clinton and Saddam were cut out of two different images and pasted on the image of the White House. Care was taken to incorporate the speaker stands with microphones, while preserving the correct shadows and illumination. Figure 1 is, in fact, an example of a very realistic counterfeiting. Another example of digital forgeries was given in the plenary talk by Dr. Tomaso A. Poggio at Electronic Imaging 2003 in Santa Clara. In his talk, Dr. Poggio showed how engineers can learn the movements of anyone's lips from a short video and then digitally manipulate the lips to arbitrarily alter spoken content. In a good example, a video segment showing a television anchor announcing evening news was altered to make the anchor appear by singing a popular song instead, while preserving the match between the sound and the motion of the lips.
