Many parallel algorithms have been proposed for finding the correct matches between feature points in random dot stereograms. Some algorithms have used local support functions and have achieved globally good solutions by using relaxation in a parallel network. Recently, Prazdny has shown that iteration is unnecessary if a much larger support function is used, and that this support function can be desiqned to work for stereograms containing transparent surfaces. We describe a simple global support function that can be efficiently implemented by relaxation in a network with only local connectivity. This function, which is the solution to the heat diffusion equation,does not work as well as Prazdny's. By using the difference of two heat equations, we can improve the performance and get results almost identical to Prazdny's at a lower computational cost.

PDF file

In  IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'85)

Publisher  IEEE Computer Society Press

Details