1,288 followers
7/ EISEN thus predicts, from a static image, pairwise affinities between scene elements. But affinities aren't objects. Inspired by neuroscience, we devise a *neural grouping primitive* to extract segments from visual affinities. (Roelfsema & Houtkam