Efficient Passive Forgery Detection in Digital Images

Abstract

The revolution in digital imaging technologies has changed our lives. Digital images are flooding in our daily life. The three main functions of an image are: conveying of information, memory sharing, and artistry. In comparison to textual information, information through an image is more direct and instant. newline In general, a digital image is generated using either the digital camera or using some computer graphics rendering software. Therefore, based on the generation process, digital images can be classified as computer-generated and real photographic images. An image created using computer graphics rendering software is known as a computer-generated image. A photographic image is an image that is captured using a digital camera. newline Image forgery detection methods can be broadly categorized into active and passive forgery detection methods. The active forgery detection method is based on additional information embedded in the digital image. However, the passive forgery detection method detects the forgery by extracting intrinsic features within the image. newline This thesis presents eight new passive approaches to solve the problem of image forgery detection in digital images. The first three approaches address the problem of detecting computer-generated images. The next three approaches are proposed to detect the copy-move forgery in an image, whereas the last two approaches are designed to detect image splicing forgery. Various concepts such as Tetrolet transform [109], Gaussian-Hermite moments [254], Neuro-fuzzy classifier [217], etc. are explored and used in the proposed approaches. Experimental validation of all the proposed methods has been performed on various available image datasets, and the results are presented in the thesis. Finally, the key findings of the thesis are concluded and the scope of the future work is provided. newline newline