Abstract
With the proliferation of a large number of digital tools and techniques in recent years, it becomes a challenge to tackle the crimes in the digital world like forgery or duplication of official documents. Forgery detection is a very difficult task in case of digital images if the source image is unavailable. Moreover, the problem becomes much more complex when it has to be detected directly in the compressed domain. Most of the existing forgery detection techniques are unable to work directly with the compressed digital image and fail to detect forgery within the compressed image. Therefore, this research paper aims to demonstrate two unsupervised algorithms for forgery detection Copy-Move and Copy-Paste based forged scenarios - directly in the JPEG compressed domain.