Robust Technique of Localizing Blurred Image Splicing Based on Exposing Blur Type Inconsistency

the process of image tampering is nothing but digital process that needs the knowledge good visual creativity as well as image properties. There are different kinds of image tampering such as copy & move, splicing, resize, cropping etc. In this paper we are focusing on blurred image splicing. The image splicing is done for many reasons, the most critical impacts of image splicing related to security and forensic analysis. Therefore it is required to have automated technique which can able to detect and localize the tampered region from input digital image accurately and efficiently. In this paper we are presenting efficient method for localizing blurred image splicing using blur type inconsistency. This method overcomes the limitations of existing framework for blurred image splicing localization depending on partial blur type inconsistency. The proposed method is based on block based processing. Image is divided into overlapping blocks, after that local blur type detection is performed by extracting the features in order to estimate the local blur kernels. The estimation is done using maximum a posteriori (MAP) technique. Using the extracted features, we further do the processing of blocks classification into motion blur (Negative class) and out-of-focus (Positive class) in order to generate invariant blur type regions. Then the technique of fine splicing location is then applied in order to increase the regions boundary precision. To overcome the problem of conflicts between out-of-focus block and motion blur block if they presented in same area, we are using fuzzy logic rules. This can improve the accuracy of slicing detection process as compared to existing.