Passive Detection of Splicing and Copy-Move Attacks in Image Forgery

Internet of Things (IoT) image sensors for surveillance and monitoring, digital cameras, smart phones and social media generate huge volume of digital images every day. Image splicing and copy-move attacks are the most common types of image forgery that can be done very easily using modern photo editing software. Recently, digital forensics has drawn much attention to detect such tampering on images. In this paper, we introduce a novel feature extraction technique, namely Sum of Relevant Inter-Cell Values (SRIV) using which we propose a passive (blind) image forgery detection method based on Discrete Cosine Transformation (DCT) and Local Binary Pattern (LBP). First, the input image is divided into non-overlapping blocks and 2D block DCT is applied to capture the changes of a tampered image in the frequency domain. Then LBP operator is applied to enhance the local changes among the neighbouring DCT coefficients, magnifying the changes in high frequency components resulting from splicing and copy-move attacks. The resulting LBP image is again divided into non-overlapping blocks. Finally, SRIV is applied on the LBP image blocks to extract features which are then fed into a Support Vector Machine (SVM) classifier to identify forged images from authentic ones. Extensive experiment on four well-known benchmark datasets of tampered images reveal the superiority of our method over recent state-of-the-art methods.