Search

The rapid development of many open source and commercial image editing software makes the authenticity of the digital images questionable. Copy-move forgery is one of the most widely used tampering techniques to create desirable objects or conceal undesirable objects in a scene. Existing techniques reported in the literature to detect such tampering aim to improve the robustness of these methods against the use of JPEG compression, blurring, noise, or other types of post processing operations. These post processing operations are frequently used with the intention to conceal tampering and reduce tampering clues. A robust method based on the color moments and other five image descriptors is proposed in this paper. The method divides the image into fixed size overlapping blocks. Clustering operation divides entire search space into smaller pieces with similar color distribution. Blocks from the tampered regions will reside within the same cluster since both copied and moved regions have similar color distributions. Five image descriptors are used to extract block features, which makes the method more robust to post processing operations. An ensemble of deep compositional pattern-producing neural networks are trained with these extracted features. Similarity among feature vectors in clusters indicates possible forged regions. Experimental results show that the proposed method can detect copy-move forgery even if an image was distorted by gamma correction, addictive white Gaussian noise, JPEG compression, or blurring. [ABSTRACT FROM AUTHOR]

In this paper, a Canny operator-based method using PCNN (Pulse-Coupled Neural Network) is proposed for color image segmentation. The coarse location information of the salient object and the background is first estimated based on the distribution of the detected key-points. An image is then over-segmented into super-pixels and their histograms are computed. The saliency of a super-pixel is obtained according to the maximum of variance ration and Shannon entropy. Color image segmentation is implemented using PCNN based on Canny operator edge detection method. The proposed method is compared with state-of-the-art methods on the widely used dataset, and the experiments show that it overall obtains more accurate results. [ABSTRACT FROM AUTHOR]