A. Massad (Germany)
Visual Interpolation, Perceptual Grouping, Tensor Voting
The integration of fragmented contours into larger units is important for different domains of visual perception. Al though a common interpolation mechanism is assumed, the shape of the curves shows considerable variations with the completion type (amodal/modal) and contextual informa tion. Long-range completions are ruled by a mediation between the Gestalt laws of good continuation and min imal path which is controlled by a parameter called GC MP contrast. In this paper, we propose the integration of such a parameter into the tensor voting model for percep tual grouping. We show under which conditions long-range connections are possible with this model and present results of this extension on typical configurations of rectilinear and curvilinear fragments.
Important Links:
Go Back
