Gaze-Dependent Tone Mapping

In this paper we model the process of temporal adaptation of the human visual system to varying luminance conditions. An eye tracker is used to capture the location of an observer’s gaze in a high dynamic range image displayed on the screen. We apply a novel technique of eye tracker data filtering to avoid flickering caused by incorrect gaze estimation. Temporary adaptation luminance is then determined in the area surrounding the gaze point. We use its value to compress the high dynamic range image and display it on the low dynamic range display. The applied tone mapping technique uses a global compression curve in which location is shifted along the luminance axis according to a value of the adaptation luminance. This technique models the natural process of adaptation occurring in the human eyes, also taking into account the time-dependent visual adaptation to dark and bright backgrounds.