A Comparative Study on the Detection of Image Forgery of Tampered Background or Foreground

Recently tampering and manipulating of images has become easier with advanced image editing applications. The process of making a composite image by combining image fragments from different images is called image splicing. One of the common types of image splicing forgery is to change the whole background of the images for different purposes. However, very few studies in the literature directly address the tampered background or foreground detection. We present a comparative study and analysis of image splicing detection methods on the detection of image forgery of tampered background or foreground. First, we create a novel dataset contains both tampered and non-tampered images. Using the dataset, we conduct a comparison of the state-of-the-art image splicing detection methods. The experimental results indicate that existing image splicing detection methods are not sufficient for detecting background or foreground forgeries, and new methods should be developed