A Novel Method to Distinguish Photorealistic Computer Generated Images from Photographic Images

摘要

The computer graphics rendering software can be used to create computer generated images so photorealistic that it becomes almost impossible to discriminate computer generated images from photographic images (generated by digital camera) by human visual system. The use of photorealistic computer generated images has brought the revolutionary changes in the film industry and video games. On the other hand, peoples with malicious intentions are using the photorealistic computer generated images in various fields such as fake journalism, forgery in academic research etc. Therefore it becomes necessary to distinguish photorealistic computer generated images from the real photographic images. In this paper a method has been proposed to distinguish photorealistic computer generated images from the real photographic images based on the Tetrolet transform and Neuro-fuzzy classifier. Based on the experimental findings it is observed that the proposed method is more effective in terms of the classification accuracy than the state-of-the-art techniques.