Radon odd radial harmonic Fourier moments in detecting cloned forgery image

Abstract

Over the past decades, due to the progress of computing abilities and easy access to powerful photo editing software, more digital photos and images are created commonplace. Unfortunately, this technology progress may also bring with a big risk of the information security. Image region cloning is a popular and simple manner to create realistic forgery images. Most relevant researches have been carried out, but the methods based on those researches are only able to detect some simple duplicated successfully. So we present a new Radon Odd Radial Harmonic Fourier Moments (RORHFMs) method. Compared with other relevant methods, this method is more robust to resist post-processed operations, such as anti-translation, anti-rotation, anti-scaling, anti-mirror operations and resisted Gaussian noise contamination. We also introduce an auxiliary circle template to slide and detect the suspicious image in order to locate the cloned region. The invariant moment features of image are extracted and analyzed by our method. Each feature of similar region is arranged orderly by Lexicographic sorting for detecting. Pearson Correlation Coefficient is applied to calculate and classify the statistical data. Then, the statistical data is searched, analyzed. At last, the original coordinates of cloned and pasted region are detected and denoted. Extensive experiments verified better robustness of RORHFMs method than other relevant methods in detecting cloned forgery.