Continuity-based and discontinuity-based segmentation in transparent and spatially segregated global motion

The mechanisms underlying the parsing of a spatial distribution of velocity vectors into two adjacent (spatially segregated) or overlapping (transpare nt) motion surfaces were examined using random dot kinematograms. Parsing m ight occur using either of two principles. Surfaces might be defined on the basis of similarity of motion vectors and then sharp perceptual boundaries drawn between different surfaces (continuity-based segmentation). Alternat ively, detection of a high gradient of direction or speed separating the mo tion surfaces might drive the process (discontinuity-based segmentation). T o establish which method is used, we examined the effect of blurring the mo tion direction gradient. In the case of a sharp direction gradient, each cl ot had one of two directions differing by 135 degrees. With a shallow gradi ent, most dots had one of two directions but the directions of the remainde r spanned the range between one motion-defined surface and the other. In th e spatial segregation case the gradient defined a central boundary separati ng two regions. In the transparent version the dots were randomly positione d. In both cases all dots moved with the same speed and existed for only tw o frames before being randomly replaced. The ability of observers to parse the motion distribution was measured in terms of their ability to discrimin ate the direction of one of the two surfaces. Performance was hardly affect ed by spreading the gradient over at least 25% of the dots (corresponding t o a 1 degrees strip in the segregation case). We conclude that detection of sharp velocity gradients is not necessary for distinguishing different mot ion surfaces. (C) 2000 Elsevier Science Ltd. All rights reserved.