Active Image Forgery Detection: State of the Art and Possible Enhancements

Tens of thousands of digital images are produced, stored, and distributed in every minute worldwide through various digital cameras, social media, and image sharing platforms. While most of images could be real and they indicate a certain source of evidence, some others could be easily tampered and cause several detriments to us. This begs the question: How we can tackle the problem of digital image forgery detection? Detection of tampering with digital images is still an active research for the image processing and computer vision community. Over the past decade, there have been vast expansions in the designing and developing of image forgery detection algorithms. All these algorithms are divided into two categories: (1) Active, and (2) Passive. Using the Active approaches, we create and embed data or information as a cipher key into the original images to protect them against a forgery. In the Passive algorithms we only investigate some local image features such as statistical anomalies, correlations and compressions to detect forgeries.