Splicing Essay

1122 Words3 Pages

Image splicing forgery technique involves composition or merging of two or more images changing the original image significantly to produce a forged image. In case images with differing background are merged then it becomes very difficult to make the borders and boundaries indiscernible. Figure 2 below shows an example of image splicing where the face of two different people is combined to form a forged image. Fig 2: Osama bin ladin’s spliced image which became viral on the net[theguardian.com, 2nd May 2011] Splicing detection is a complex problem whereby the composite regions are investigated by a variety of methods. The presence of abrupt changes between different regions that are combined and their backgrounds, provide valuable traces to detect splicing in the image under consideration. Farid [56] suggested a method based on bi-spectral analysis to detect introduction of un-natural higher-order correlations into the signal by the forgery process and is successfully implemented for detecting human-speech splicing. Ng and Chang [57] suggested an image-splicing detection method based on the use of bi-coherence magnitude features and phase features. Detection accuracy of 70% was obtained. Same authors later developed a model for detection of discontinuity caused by abrupt splicing using bi-coherence [58]. Fu et al. [59] proposed a method that implemented use of Hilbert-Huang transform (HHT) to obtain features for classification. Statistical natural image model defined by moments of characteristic functions was used to differentiate the spliced images from the original images. Chen et al. [60] proposed a method that obtains image features from moments of wavelet characteristic and 2-D phase congruency which is a sensitive m... ... middle of paper ... ...ction for the tampering decision. Texture- and edge-based features are extracted from illuminant estimators which are then provided to a machine-learning approach for automatic decision-making. SVM is used for classification and detection rates of 86% on a dataset consisting of 200 images and 83% on 50 images collected from the Internet was achieved. Another detection scheme based on blur as a clue is proposed in [77]. This method expose the presence of splicing by evaluating inconsistencies in motion blur even under space-variant blurring situations. The methods discussed above have a few limitations such as the detection methods fail when measures such as blur are used to conceal the sharp edges disturbances after splicing. The requirement of edges to be wide for reliable extraction is also a limitation. Moreover minor and localized tampering may go undetected.

More about Splicing Essay

Open Document