Robust detection of median filtering based on data-pair histogram feature and local configuration pattern

Abstract

Median filtering is a widely used method for removing noise, preserving edges and smoothing regions within an image. Detection of median filtering has drawn much attention from researchers of image forensic. A new robust detection scheme of median filtering based on combined features of difference image is proposed in this paper. In the proposed approach, the combined features consist of data-pair histogram features (DPHFs) and Local configuration pattern (LCP) of first-order difference image (DI). A large number of experiments on single dataset and compound datasets show that, the proposed scheme not only achieves super performance on the uncompressed image datasets, but also achieves better detection performance compared with state-of-the-art methods, especially for strongly JPEG compressed and low resolution images.