Image Forgery Detection using Modified Adaptive Over-Segmentation and Feature Matching
With the emergence of digitisation of Visual Graphics it has been a constant need to extract typical features of Images. Forgery, on the other hand, is not something new to the world of Computer Graphics. In the past it was restricted to craftsmanship and writing and it didn’t affect human civilisation rather it was seen as an art for a few centuries. But, with the advent of the technology it has been noticed that by the use of computerised picture handling softwares and altering devices a picture can be easily changed and controlled. It is notably done nowadays to falsify and spread certain narratives through the Media and on the Internet. These alterations show that even graphical information is as vulnerable as any other information, questioning the credibility of digital images. This work is an improvement of an already existing Adaptive Over- Segmentation technique which has recently been used to detect forgery of a digital image by introducing a new feature of detection of forgery even if noise was introduced to the forged copy of the original image.