Imperfect Understandings: A Grounded Theory And Eye Gaze Investigation Of Human Perceptions Of Manipulated And Unmanipulated Digital Images

To investigate the extent to which humans are able to perceive manipulated images, and how they interpret these images, we use eye gaze tracking and grounded theory analysis of 80 participants viewing a combination of manipulated and unmanipulated images together with pre-familiarisation with image manipulation techniques. We find that accurate identification of manipulations is poor to moderate, and ability to identify what has been manipulated is poor. However, extended visual attention to manipulated regions was associated with greater accuracy. Grounded theory analysis shows that uncertainty in respect of veracity of images viewed is common across all images. Similarly, the use of logic influences participants success rate in identifying manipulations.