| D.A. Forsyth, J. Haddon, and S. Ioffe, "The Joy of Sampling," Int'l J. Computer Vision, vol. 41, no. 1/2, pp. 109-134, 2001. |
....minima, or as statistical inferences based on fair sampling or expectation value integrals over highly multi modal distributions. Importance sampling is a promising approach for such applications, particularly when combined with sequential ( Markov Chain Monte Carlo ) layered or annealed samplers [8, 4, 5], optionally punctuated with bursts of local optimization [10, 3, 25] Sampling methods are flexible, but they tend to be computationally expensive for a given level of accuracy. In particular, when used on multi modal cost surfaces, current sequential samplers are very prone to becoming trapped ....
....sequences using an edge and intensity based cost function, with a combination of robust constraint consistent local optimization and oversized covariance scaled sampling to focus samples on probable low cost regions. Hyperdynamics uses stochastic dynamics with cost gradient based sampling as in [8, 17, 4], but boosts the dynamics with a novel importance sampler constructed from the original probability surface using local gradient and curvature information. All of the annealing methods try to increase transition rates by sampling a modified distribution, but only the one given here specifically ....
D. Forsyth, J. Haddon, and S. Ioffe. The Joy of Sampling. IJCV, 41:109--134, 2001.
No context found.
D.A. Forsyth, J. Haddon, and S. Ioffe, "The Joy of Sampling," Int'l J. Computer Vision, vol. 41, no. 1/2, pp. 109-134, 2001.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC