부산대학교 도서관

Altering an original picture in order to create fake image by means of any fraud or other reason is known to be image forgery. As per the process involved in fake image creation, image forgery scenarios can be divided into three groups --- Image retouching, image splicing and copy-move attack. Numerous types of techniques of image forensics are used to detect forgery. These types can be helpful in active and passive protection. There are many existing techniques or approaches used to detect the forgery in an image but provides less accurate result. The proposed work here deals with image forgery detection using rotation invariant technique with blur-invariant copy-move forgery detection. The proposed mechanism firstly rotates the inbuilt object of an image and then divides the image into small blocks. After dividing the image into small sized blocks, fast Fourier transformation is applied and then detected. The simulations are performed in MATLAB tool. The method is more accurate and provide accuracy rate between 90-95%.