In the present world of social networking, image manipulation is the easiest and the scariest job! This requires an utter need of efficient image manipulation detection techniques. Moving from the traditional image manipulation detection techniques to the present scenario, CNN can be the perfect deep learning model as the human visual cortex has the ability to detect tampered areas in an image!
With the advent of social networking services like Facebook and Instagram in the past few decades, the sharing of digital images has substantially increased. Also, these images can be easily modified these days by using easily available image processing softwares like Adobe Photoshop. These modified images are used for fake news, mob incitement, etc. Hence there is a crucial need for image authentication schemes that can verify if the image is authentic or manipulated. Earlier, most of the research in this direction has been done in pixel based, format based, camera based, physics based and geometry based schemes. All these schemes work upon the visual information of the image. Recently, CNN’s came into picture that are inspired by visual cortex. It was analyzed that it is possible for a human visual cortex to detect tampered areas in an image. Thus CNN can be the perfect deep learning model for this job! Content to be discussed: Traditional Image Manipulation Detection Techniques Various datasets available for Image Manipulation Detection Experimentation Learning Rich Features for Image Manipulation Detection CNN based Image Manipulation Detection Techniques