ABSTRACT
In today's cyber world, digital multimedia such as images and videos act as the most frequently transmitted information carriers. Also, due to rapid growth in importance and impact of digital evidence worldwide, digital multimedia form the major source of evidence for any event in media, broadcast, and legal industries. However, digital multimedia, such as images and videos, can be easily manipulated using widely available image and video editing tools. Hence, any given digital image and video cannot be trusted blindly as evidence to events/circumstances. Manipulated multimedia is capable of adversely impacting political views and law and order of an entire nation. So, multimedia modification attack is a crucial problem in cyber space. To address this problem, deep convolutional neural network (CNN) model–based digital forensic measures have been presented in this chapter to verify the integrity and authenticity of digital images. The presented deep learning–based forensic framework efficiently detects and localizes the forged region(s) in double-compressed JPEG images. The experimental results establish the efficiency of the proposed model.
