Christian Richardt, Lech Swirski, Ian P. Davies, Neil A. Dodgson
We introduce a novel computational model for objectively assessing the visual comfort of stereoscopic 3D imagery. Our model integrates research in visual perception with tools from stereo computer vision to quantify the degree of stereo coherence between both stereo half-images. We show that the coherence scores computed by our model strongly correlate with human comfort ratings using a perceptual study of 20 participants rating 80 images each. Based on our experiments, we further propose a taxonomy of stereo coherence issues which affect viewing comfort, and propose a set of computational tools that extend our model to identify and localise stereo coherence issues from stereoscopic 3D images.
Categories and Subject Descriptors (according to ACM CCS): I.2.10 [Artificial Intelligence]: Vision and Scene Understanding-Perceptual reasoning; I.4.8 [Image Processing]: Scene Analysis-Depth cues & stereo