CIE x050-OP076 PDF

An objective evaluation of the uniformity of luminance distributions has been of strong interest in recent years. While it takes no effort for humans to detect profound non-uniformities, the development of an objective measurement method independent of application has posed to be a great challenge.

Recently the authors proposed a functional model that mimics the first contrast processing steps of the human visual system to predict the visibility of non-uniformities. However, this model does not regard stimulus area dependent effects which are known to significantly influence the contrast threshold. A divisive normalization stage is added to overcome this problem.

The simulated model results are in accordance with psychophysical measurements of periodical stimuli but show differences to measured data for stimuli with a sharp and closed contour. A change in model structure as well as further enhancement is needed for the goal of evaluating luminance distributions of arbitrary shape.