Search

Cloning is one of the most used software tools for unauthorized changes to digital images. Therefore, the task of identifying cloning is actual. To increase the information value of cloning detec-tion results it is necessary to separate the clone from the prototype, it is necessary to determine which of the obtained regions is a clone. If the postprocessing of the clone and the prototype is the same, then this task is the most difficult. We will assume that the clone and prototype are found in the digital image. The method of separation of the clone from the prototype has been developed. It is based on theoretical propositions suggested by the authors earlier. The method is effective in the absence of post-processing of the image. We consider pairs of corresponding blocks from the regions of the clone and the prototype. А distinctive neighborhood of radius 1 is constructed for each block. The distinctive neighborhood reflects the cumulative difference of the block from the nearest neighboring blocks. The norm of the distinctive neighborhood of radius 1 is the main quantitative indicator that allows us to separate the clone from the prototype. The algorithm that implements the developed method is proposed. The efficiency of the algo-rithm exceeds the efficiency of existing analogues. The results of the computational experiment are given.

One of the widely and often used for falsifications of digital images of software tools remains cloning. Today much attention is paid to the task of revealing cloning, but the question of determining which of the identified areas of the image is a clone/prototype remains. There are different approaches that are used to solve this problem. One of the main approaches is the use of digital watermark technology. But embedding of digital watermarks in the image violates its integrity. It is undesirable. Another approach is to search for the results of the post-processing of the clone, which differs from the processing of the prototype. But in the case when the clone/prototype are small, clone postprocessing is often absent. This makes the methods of this group incompetent in this case. Previously the authors developed a method for separating the clone from the prototype in the absence of additional processing of the cloned digital image. This article proposes an improvement of this method. The goal of the improvement is to increase the informative value of cloning detection results in a digital image. The improved method is effective in conditions of additional perturbing effects, in particular, compression with losses of the cloned image. This method does not use digital watermark technology. Improvement of the method is provided by detecting and fixing the moment of the jump in the value of the modulus of the difference in the norms of the distinctive neighborhoods of the corresponding clone and prototype blocks. The distinctive neighborhood of the block B is (2k + 1)*(2k + 1)-matrix. Elements of this matrix reflect the difference of B from the image blocks, which are from B at a distance not exceeding k. This jump takes place when these blocks reach the boundaries of the real regions of the clone and the prototype. At the moment of fixing the jump, the clone is answered by a block whose norm of a distinctive neighborhood is larger. The process of fixing the jump occurs for several (max. 8) pairs of corresponding blocks. The final conclusion is made taking into account which of the identified areas was defined as a clone/prototype more times. The results of the computational experiment indicate the effectiveness of the proposed method in the presence of additional perturbing effects on the cloned digital image. [ABSTRACT FROM AUTHOR]

The cloning is one of the most common and often used software tool for unauthorized changes of digital images that implemented in all modern graphics editors. Thus, the task of detection the results of cloning is one of the most urgent for verifying the integrity of digital content. This work is devoted to increasing the informative value of the results of the detection of integrity violations of the digital images due to the development of a new method for separating of clone and prototype in the absence of their post-processing or the absence of differences in this treatment. The developed method does not use the technology of digital watermarks which leads to a violation of the integrity of the original image. The effectiveness of the algorithmic implementation of the developed method is practically independent of the value of the disturbing effect used in the post-processing of the cloned image, and increases with decreasing absolute sizes of the clone and pre-image. Analogues of the proposed method from the point of view of the problem being solved (separating of clone and prototype) and the proposed conditions have not been found in the open sources. [ABSTRACT FROM AUTHOR]