Daniel Keren, "Probabilistic analyses of interpolation in computer vision," Ph.D. thesis, Hebrew University of Jerusalem, 1990. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
A Full Bayesian Approach to Curve and Surface Reconstruction - Keren, Werman (1999) (3 citations) Self-citation (Keren) (Correct)
No context found.
Daniel Keren, "Probabilistic analyses of interpolation in computer vision," Ph.D. thesis, Hebrew University of Jerusalem, 1990.
A Bayesian Framework for Regularization - Keren, Werman (1994) (1 citation) Self-citation (Keren) (Correct)
.... The question of how to compute such integrals s those ppering in Equation 1 which re defined over domains that re infinite dimensional hs been solved for some types of integrals in the realm of pure mthemtics [8, 14, 16, 29, 15, 7] It ws pplied to the types of spces used in regulriztion in [12, 13]. The spce M, is Hilbert spce [30] Let us recall that if U is subspce of Hilbert spce H, its orthwonal subspace, U , is defined s following 4 It is well known that for every h H T there are u I U and u2 U so that Ul u2 = h. Also T Ul and u2 are unique. They are called the projections of u ....
Daniel Keren. Probabilistic Analyses of Interpolation in Computer Vision. PhD thesis, Hebrew University of Jerusalem, 1990.
A Full Bayesian Approach to Curve and Surface Reconstruction - Keren, Werman (1999) (3 citations) Self-citation (Keren) (Correct)
....the equality ( f, G x,y ) 2D = f (x, y) where ( f, g) 2D = ## (f uu g uu 2 f uv g uv f vv g vv ) dudv. As opposed to the one dimensional reproducing kernels, which have a simple form (cubic splines) there is no known closed form expression for the 2D reproducing kernels. In [15, 16] this problem is addressed, and it is shown how to quickly compute the functions G x,y (u,v) to any desired accuracy using an approximation on a finite subspace. Due to the fact that the inner product ( f, g) 2D is nearly orthogonal on subspaces spanned by trigonometric functions, convergence is ....
....computed using Eq. 9) Figure 12. Two nearly identical data sets superimposed. Figure 13. The (scaled) probability distribution for one of the data sets of Fig. 12. distribution determined by a large # are very much alike , in the sense that random samples from these spaces are very similar [16]. It makes sense therefore to use the average smoothness of the functions to determine the prior. This average smoothness is # ( # f 2 uu du)exp( # # f 2 uu du)D f . The onedimensional equivalent is # x 2 exp( #x 2 ) dx, which equals # 3 2 . 38 Keren and Werman The prior, ....
[Article contains additional citation context not shown here]
Daniel Keren, "Probabilistic analyses of interpolation in computer vision," Ph.D. thesis, Hebrew University of Jerusalem, 1990.
A Bayesian Framework for Regularization - Keren, Werman (1994) (1 citation) Self-citation (Keren) (Correct)
.... of how to compute such integrals as those appearing in Equation 1 which are defined over domains that are infinite dimensional has been solved for some types of integrals in the realm of pure mathematics [8, 14, 16, 29, 15, 7] It was applied to the types of spaces used in regularization in [12, 13]. The space M ;oe is a Hilbert space [30] Let us recall that if U is a subspace of a Hilbert space H , its orthogonal subspace, U , is defined as following U = fh 2 H ju 2 U = u; h) 0g It is well known that for every h 2 H , there are u 1 2 U and u 2 2 U so that u 1 u ....
Daniel Keren. Probabilistic Analyses of Interpolation in Computer Vision. PhD thesis, Hebrew University of Jerusalem, 1990.
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