Interpolation processes in the visual perception of objects

• Authors: P. J. Kellman
• Venue: Neural Networks - 2003 Special issue: Advances in neural networks research — IJCNN'03
• Year: 2003

Visual perception of objects depends on segmentation and grouping processes that act on fragmentary input. This paper gives a brief overview of these processes. A simple geometry accounting for contour interpolation is described, and its applications to 2D, 3D, and spatiotemporal object interpolation processes are considered. A method is described for distinguishing interpolation based on this geometry from more global or top-down influences. Results suggest a separation between interpolation based on relatively local stimulus relations, which give rise to precise boundary representations, and processes involving recognition from partial information, which do not. Aspects of the model--especially the unified treatment of illusory and occluded objects--raise questions about the nature of seeing. Although it is often believed that illusory objects are perceived, while occluded objects are inferred, I suggest that both research and theory converge in supporting a more unified account. Illusory and occluded contours and surfaces do not divide into the real, the perceived, and the inferred, but are all represented, and in key respects, derive from identical perceptual processes.