Illuminant Color Inconsistency as a Powerful Clue for Detecting Digital Image Forgery: A Survey

摘要

Digital images capture our attention and are retained in our memory for longer than other sensory perceptions. Despite numerous instances of image forgery, still, people tend to believe digital images. At the same time, digital investigations reveal an increasing trend of image forgery with illicit purposes. Image editing operations that lead to forgery always leave traces. Investigators rely upon these traces for detecting an image forgery. Researchers are trying to detect image forgery by devising techniques that exploit the traces present in forged images. Recently, illuminant color, the color of the scene illumination present in the image that hints the illumination prevailed at the time of image capture is studied as potential evidence for image forgery. In this survey, we explore the evolution of illuminant color based image forgery detection. This survey provides a brief description of different illuminant color estimation approaches employed in image forgery detection followed by a detailed review of existing illuminant color inconsistency based forgery detection techniques. The major contribution of this survey is the elaborate discussion of future research directions to provide insight to researchers.