DocumentCode
2560703
Title
Analytic solution of stochastic completion fields
Author
Thornber, K.K. ; Williams, L.R.
Author_Institution
NEC Res. Inst., Princeton, NJ, USA
fYear
1995
fDate
21-23 Nov 1995
Firstpage
617
Lastpage
622
Abstract
Uses generalized particle trajectories to derive an analytic expression characterizing the probability distribution of boundary-completion shape. This is essential to the understanding of the perceptual phenomenon of illusory (subjective) contours. The particles´ dynamics include Poisson-distributed ensembles of driving forces as well as particle decay. The resulting field, representing completed surface boundaries, is characterized by the fraction of particles at x with velocity x˙. The distributions are projectively covariant in the sense that fields calculated in any lower-dimensional projection correspond to the projections of fields calculated in any higher dimension. Being analytic, the relationship between velocity, diffusivity, and decay can be made readily apparent
Keywords
Monte Carlo methods; computer vision; integration; probability; stochastic processes; Poisson-distributed ensembles; boundary-completion shape; driving forces; generalized particle trajectories; illusory contours; particle decay; probability distribution; stochastic completion fields; Analysis of variance; Computer vision; Distributed computing; Humans; Jacobian matrices; National electric code; Probability distribution; Psychology; Shape; Stochastic processes;
fLanguage
English
Publisher
ieee
Conference_Titel
Computer Vision, 1995. Proceedings., International Symposium on
Conference_Location
Coral Gables, FL
Print_ISBN
0-8186-7190-4
Type
conf
DOI
10.1109/ISCV.1995.477070
Filename
477070
Link To Document