Robust Image Forgery Detection Using Point Feature Analysis

Abstract

Day for day it becomes easier to temper digital images. Thus, people are in need of various forgery image detection. In this paper, we present forgery image detection techniques for two of the most common image tampering techniques; copy-move and splicing. We use match points technique after feature extraction process using SIFT and SURF. For splicing detection, we extracted the edges of the integral images of $Y, C_{b}$, and $C_{r}$ image components. GLCM is applied for each edge integral image and the feature vector is formed. The feature vector is then fed to a SVM classifier. For the copy-move, the results show that SURF feature extraction can be more efficient than SIFT, where we achieved 80% accuracy of detecting tempered images. On the other hand, processing the image in $YC _{b}C_{r}$ color model is found to give promising results in splicing image detection. We have achieved 99% true positive rate for detecting splicing images.