The Recent Digital Image Splicing Detection

Abstract

With the great progress of multimedia, digital image tampering has become a critical challenge to modern social credit. Among these abundant tampering methods, image splicing is especially noteworthy. This paper is based on the search and organization of classical and recent papers, aiming to provide a taxonomy analysis on this topic. By classifying the image splicing detection technology into two main classes, four different sub-classes are proposed. Specifically, depending on whether artificial intelligence is applied, traditional methods and deep learning methods are divided from each other. A further classification of the traditional methods can be proposed based on the specific aspect of anomaly focused on by each method. Also, this paper is expected to provide a channel for researchers to quickly get to know the digital image splicing detection area. Besides, this paper finally hopes to provide a way to help researchers understand the status of this area and to conduct their technical assessments.