Study of Image Splicing Detection

摘要

With the advent of digital technology, digital image has gradually taken the place of the original analog photograph, and the forgery of digital image has become more and more easy and indiscoverable. Image splicing is a commonly used technique in image tampering. In this paper, we simply introduce the definition of image splicing and some methods of image splicing detection, mainly including the detection based on steganalysis model, the detection based on Hilbert-Huang transform (HHT) and moments of characteristic functions (CF) with wavelet decomposition. We focus on discussing our proposed approach based on image quality metrics (IQMs) and moment features. Especially we analyze the model creation and the extraction of features in digital image. In addition, we compare these approaches and analyze the future works of digital image forensics.