Image splicing detection based on moment features and Hilbert-Huang Transform

Image splicing is considered to be more simple and common than other image tamper technology. And the image splicing problem can be tackled as a two-class classification problem under the pattern recognition framework. By analyzing the different characteristics between real images and spliced images, an image splicing blind detection approach based on moment features and Hilbert-Huang Transform is proposed. This algorithm extracts two groups of features from the first order histogram of the image DWT (Discrete Wavelet Transform) coefficients and the Hilbert-Huang Transform of the image. Then the features are fed into the Support Vector Machine to classify real images and spliced images. Experimental results show that the average accuracy rate can achieve 85.8696%, and the real-time is improved.