Image splicing forgery detection using local binary pattern and discrete wavelet transform

A common way of tampering in digital images is known as splicing, Where a selected region from an image is pasted into another or same image. In this paper we propose a method to detect image manipulation. Firstly the algorithm converts input RGB image into YCbCr color channel, afterwards chrominance component is divided into non-overlapping blocks. Secondly Local Binary Pattern (LBP) operator is performed, and wavelet transform is applied in all blocks. Finally, Principal Component Analysis (PCA) is used for all blocks and the output is fed to Support Vector Machine (SVM) classifier as features. Experimental results demonstrate the efficiency of proposed method in exposing image splicing forgeries.