Hybrid Feature Classification Model with Probabilistic Classification of the Image Forgery Detection

The image forgery detection is the technique to find the forgery in the images by analyzing the image matrix against the ground truth image. The image forgery detection plays the very important role for the digital media houses. The image data theft or copyright forgery may cause the hefty monetary losses to the real owners of the digital image data. To protect the copyright of the real owner, the image forgery methods are incorporated over the suspected image data. Several factors are analyzed under the proposed model in order to recognize the real identity of the image, which is decided in the terms of authentic or forged image. In this paper, the proposed model has been designed with the multiple features based image forgery evaluation. The incorporation of the fast retina key points (FREAK) along with the speeded up robust features (SURF) have been utilized for the development of the robust feature descriptor. The support vector machine (SVM) has been used for the probabilistic classification of the model. The experimental results have been obtained in the form of statistical parameter after the testing under the various numbers of test cases. The experimental results have clearly defined the proposed model as the winner in comparison with the existing models.