The Experts below are selected from a list of 297 Experts worldwide ranked by ideXlab platform

Pier Luigi Dragotti - One of the best experts on this subject based on the ideXlab platform.

  • Quadtree structured restoration algorithms for Piecewise Polynomial images
    2009 IEEE International Conference on Acoustics Speech and Signal Processing, 2009
    Co-Authors: Adam Scholefield, Pier Luigi Dragotti
    Abstract:

    Iterative shrinkage of sparse and redundant representations are at the heart of many state of the art denoising and deconvolution algorithms. They assume the signal is well approximated by a few elements from an overcomplete basis of a linear space. If one instead selects the elements from a nonlinear manifold it is possible to more efficiently represent Piecewise Polynomial signals. This suggests that image restoration algorithms based around nonlinear transformations could provide better results for this class of signals. This paper uses iterative shrinkage ideas and a nonlinear quadtree decomposition to develop image restoration algorithms suitable for Piecewise Polynomial images.

  • ICASSP - Quadtree structured restoration algorithms for Piecewise Polynomial images
    2009 IEEE International Conference on Acoustics Speech and Signal Processing, 2009
    Co-Authors: Adam Scholefield, Pier Luigi Dragotti
    Abstract:

    Iterative shrinkage of sparse and redundant representations are at the heart of many state of the art denoising and deconvolution algorithms. They assume the signal is well approximated by a few elements from an overcomplete basis of a linear space. If one instead selects the elements from a nonlinear manifold it is possible to more efficiently represent Piecewise Polynomial signals. This suggests that image restoration algorithms based around nonlinear transformations could provide better results for this class of signals. This paper uses iterative shrinkage ideas and a nonlinear quadtree decomposition to develop image restoration algorithms suitable for Piecewise Polynomial images.

M. Rochdi - One of the best experts on this subject based on the ideXlab platform.

  • The Case of Piecewise-Polynomial Static-Gain
    2020
    Co-Authors: F. Girl, M. Rochdi
    Abstract:

    We are considering the problem of identifying Hammerstein models, including a nonlinear static gain in series with linear dynamics. We particularly focus on static gains that are Piecewise Polynomial within a finite interval. For this class of models we develop an identification scheme with the following main features: i) a dynamics-oriented parameterization and a static-gain- oriented parameterization, ii) a parameter estimation algorithm associated with each parameterization, iii) a persistently exciting input signal. The parameter estimates thus obtained are shown to converge to their true values, yielding asymptotically the true model.

  • Hammerstein model identification: the case of Piecewise-Polynomial static-gain
    Proceedings of the 2001 American Control Conference. (Cat. No.01CH37148), 2001
    Co-Authors: F. Giri, F.z. Chaoui, M. Rochdi
    Abstract:

    We are considering the problem of identifying Hammerstein models, including a nonlinear static gain in series with linear dynamics. We particularly focus on static gains that are Piecewise Polynomial within a finite interval. For this class of models we develop an identification scheme with the following main features: (i) a dynamics-oriented parameterization and a static-gain-oriented parameterization, (ii) a parameter estimation algorithm associated with each parameterization, (iii) a persistently exciting input signal. The parameter estimates thus obtained are shown to converge to their true values, yielding asymptotically the true model.

Adam Scholefield - One of the best experts on this subject based on the ideXlab platform.

  • Quadtree structured restoration algorithms for Piecewise Polynomial images
    2009 IEEE International Conference on Acoustics Speech and Signal Processing, 2009
    Co-Authors: Adam Scholefield, Pier Luigi Dragotti
    Abstract:

    Iterative shrinkage of sparse and redundant representations are at the heart of many state of the art denoising and deconvolution algorithms. They assume the signal is well approximated by a few elements from an overcomplete basis of a linear space. If one instead selects the elements from a nonlinear manifold it is possible to more efficiently represent Piecewise Polynomial signals. This suggests that image restoration algorithms based around nonlinear transformations could provide better results for this class of signals. This paper uses iterative shrinkage ideas and a nonlinear quadtree decomposition to develop image restoration algorithms suitable for Piecewise Polynomial images.

  • ICASSP - Quadtree structured restoration algorithms for Piecewise Polynomial images
    2009 IEEE International Conference on Acoustics Speech and Signal Processing, 2009
    Co-Authors: Adam Scholefield, Pier Luigi Dragotti
    Abstract:

    Iterative shrinkage of sparse and redundant representations are at the heart of many state of the art denoising and deconvolution algorithms. They assume the signal is well approximated by a few elements from an overcomplete basis of a linear space. If one instead selects the elements from a nonlinear manifold it is possible to more efficiently represent Piecewise Polynomial signals. This suggests that image restoration algorithms based around nonlinear transformations could provide better results for this class of signals. This paper uses iterative shrinkage ideas and a nonlinear quadtree decomposition to develop image restoration algorithms suitable for Piecewise Polynomial images.

Marie Sauve - One of the best experts on this subject based on the ideXlab platform.

  • Piecewise Polynomial Estimation of a Regression Function
    IEEE Transactions on Information Theory, 2010
    Co-Authors: Marie Sauve
    Abstract:

    We deal with the problem of choosing a Piecewise Polynomial estimator of a regression function s mapping [0,1] p into R. In a first part of this paper, we consider some collection of Piecewise Polynomial models. Each model is defined by a partition M of [0,1] p and a series of degrees d = (d J)J¿M ¿ NM. We propose a penalized least squares criterion which selects a model whose associated Piecewise Polynomial estimator performs approximately as well as the best one, in the sense that its quadratic risk is close to the infimum of the risks. The risk bound we provide is nonasymptotic. In a second part, we apply this result to tree-structured collections of partitions, which look like the one constructed in the first step of the CART algorithm. And we propose an extension of the CART algorithm to build a Piecewise Polynomial estimator of a regression function.

Ron Meir - One of the best experts on this subject based on the ideXlab platform.

  • almost linear vc dimension bounds for Piecewise Polynomial networks
    Neural Information Processing Systems, 1998
    Co-Authors: Peter L Bartlett, Vitaly Maiorov, Ron Meir
    Abstract:

    We compute upper and lower bounds on the VC dimension and pseudodimension of feedforward neural networks composed of Piecewise Polynomial activation functions. We show that if the number of layers is fixed, then the VC dimension and pseudo-dimension grow as W log W, where W is the number of parameters in the network. This result stands in opposition to the case where the number of layers is unbounded, in which case the VC dimension and pseudo-dimension grow as W2 . We combine our results with recently established approximation error rates and determine error bounds for the problem of regression estimation by Piecewise Polynomial networks with unbounded weights.

  • NIPS - Almost Linear VC Dimension Bounds for Piecewise Polynomial Networks
    1998
    Co-Authors: Peter L Bartlett, Vitaly Maiorov, Ron Meir
    Abstract:

    We compute upper and lower bounds on the VC dimension and pseudodimension of feedforward neural networks composed of Piecewise Polynomial activation functions. We show that if the number of layers is fixed, then the VC dimension and pseudo-dimension grow as W log W, where W is the number of parameters in the network. This result stands in opposition to the case where the number of layers is unbounded, in which case the VC dimension and pseudo-dimension grow as W2 . We combine our results with recently established approximation error rates and determine error bounds for the problem of regression estimation by Piecewise Polynomial networks with unbounded weights.