Digital Image forgery Detection by Utilize combined Feature extraction techniques

Due to the increased revolution of digital technology, the process of information sharing, accessing becomes easier. But securing this information is the major critical task. The major threat is occurred in digital images by making forgeries. Several existing techniques are utilized for detection the forgeries in digital images. But still, it lacks inaccurate detection. Hence a novel technique is designed for detecting the forged images accurately. The main motive of this research is focused on detect image forgery and localize the forged region accurately. Initially, the input images obtained from digital image acquisition and the selected images are isolated as an overlapping patch. Polar Cosine Transform (PCT) with orthogonal kernel and Local Binary Pattern (LBP) approaches are used to extract features from these patches. From the features extracted from the PCT approach, the patches are detected using Multidimensional Spectral Hashing techniques (MSH) and the forged patches are filtered out. Alternatively, geometry-based image forgery detection is carried out using the LBP extracted features. Finally, the forged regions are located and detected in the digital image. The proposed approach's efficiency is measured and compared to current techniques