Proceedings of the 2nd International Conference on Computer Science and Electronics Engineering

Image Splicing Detection Based on Statistical Properties of Benford Model

Senfeng Tong, Zhen Zhang, Yongjie Xie, Xiaodi Wu
Corresponding author
image splicing detection, statistical characteristics, AC coefficient, most significant digit, authenticity
For the operation of the image splicing tamper, this paper proposes a novel approach for blind image splicing detection based on statistical properties of the Benford model. Having made discrete wavelet transform (DWT) to the test image, we extracted alternating current (AC) coefficients in discrete cosine transform (DCT) domain of the three RGB color channels from each wavelet component, and then we calculated probability distribution of the most significant digit (MSD) of the AC coefficients in DCT domain by using the Benford model, constructed the detection model of the proposed algorithm. The threshold is set according to significant level of the statistical property difference tampered between before and after, then judgement the authenticity of the images. Experimental results show that the proposed algorithm is capable to detect splicing image efficiently.
Download article (PDF)