IMAGE SPLICING FORGERY DETECTION

Abstract

The rapid development of digital technology has raised challenges for ensuring authenticity of digital images. Image splicing, which has constituted a menace to integrity and authenticity of images, is a very common and simple trick in image tampering. Image splicing creates a composite image by cropping and pasting regions from the same or different images. Spliced images could be so eye-deceiving that they are scarcely distinguished from authentic ones even without any post processing. Therefore, image-splicing detection is of great importance in digital forensics. The forged picture however leaves some clues which can be used to locate the manipulated regions. In this paper, an effective algorithm for revealing image-splicing forgery is proposed. Firstly the algorithm converts input RGB image into YCbCr color channel. The four-level Discrete wavelet transform on each color channel is applied, the sharp edges, which are traces of cut-paste manipulation, are high frequencies and detected from LH, HL and HH sub-bands. The four level inverse discrete wavelet transform is applied by removing low frequency components and considering only high frequency components. Afterwards difference between obtained chroma components and luminance component is calculated. The morphological operation is applied to reconstruct the boundaries of sharp edge regions. The obtained largest and second largest regions are compared to determine whether the image is forged or not based on noise inconsistency level. If the obtained difference is above 1dB the input image is forged image else original. If the image is forged one, the method defines the region having maximum area to be forged region. When a fake is confirmed, suspicious regions becomes objects to be considered. Then the quality measures of the obtained result has been evaluated by means of sensitivity, specificity and accuracy. The experimental results demonstrate the robustness of the algorithm in exposing image splicing forgeries with accuracy of 88%.