Identifying Image Authenticity by Detecting Inconsistency in Light Source Direction

Inconsistency in light source direction resulted from image forgery is one of the traces which cannot be removed. Hereby, a blind identification to distinguish the image authenticity by detecting inconsistency in light source direction was proposed. The search means and least-squares method were used to calculate the light source direction for local and infinite light source images respectively, in accordance with the error function between the actual intensity and calculated intensity, as well as the constraint function suitable for different light source models. And, by analyzing if the light source directions between different objects and background in one image are consistent to determine whether the image had been altered. Experimental results showed that the correct detection rates of blind identification for local and infinite light source images were 87.31% and 85.47% respectively.