An Efficient Algorithm For Blind Image Forgery

Abstract

With the development of technology, more and more image editing tools are applied, by which images can be manipulated easily, such as resizing, cropping, copying,…. Therefore, in recent years, determining whether an image has undergone processing or not is one of the topics of great interest, and copy-move attacks are one of the most common types of image spoofing. By using modern software and tools like Photoshop, GIMP, Canva, ..., users can create multiple copy-move fake images. The thesis researches an efficient algorithm for Image Forgery Detection (IFD) based on the Scale Invariant Feature Transform (SIFT) algorithm to detect tamped origins in the image. In addition, I also built a Graphical User Interface (GUI) to make it easy for users to use this new algorithm. Although copy-move detection is research field that has been studied for more than a decade and SIFT is also applied as an effective solution to this problem, each implementation still has certain limitations. Thesis implements the SIFT algorithm to the actual image dataset, with improved accuracy with easily accessible GUI. However, due to limited time, building a private image dataset and developing a deep learning model is proposed as a further development of the topic.