no code implementations • 29 Sep 2021 • Rishi Sonthalia, Raj Rao Nadakuditi
In fact the generalization error versus number of of training data points is a double descent curve.
no code implementations • 22 May 2019 • Arvind Prasadan, Raj Rao Nadakuditi, Debashis Paul
Singular value decomposition (SVD) based principal component analysis (PCA) breaks down in the high-dimensional and limited sample size regime below a certain critical eigen-SNR that depends on the dimensionality of the system and the number of samples.
no code implementations • 5 May 2019 • Hao Wu, Raj Rao Nadakuditi
We describe a method for unmixing mixtures of freely independent random variables in a manner analogous to the independent component analysis (ICA) based method for unmixing independent random variables from their additive mixtures.
1 code implementation • 4 Mar 2019 • Arvind Prasadan, Raj Rao Nadakuditi
We show that when the latent time series are uncorrelated at a lag of one time-step then, in the large sample limit, the recovered dynamic modes will approximate, up to a column-wise normalization, the columns of the mixing matrix.
no code implementations • 6 Sep 2018 • Brian E. Moore, Saiprasad Ravishankar, Raj Rao Nadakuditi, Jeffrey A. Fessler
Sparsity and low-rank models have been popular for reconstructing images and videos from limited or corrupted measurements.
no code implementations • 18 Dec 2017 • Brian E. Moore, Chen Gao, Raj Rao Nadakuditi
We perform extensive numerical experiments on both static and moving camera video subject to a variety of dense and sparse corruptions.
no code implementations • 24 Oct 2017 • Andrew J. Wagenmaker, Brian E. Moore, Raj Rao Nadakuditi
We propose a model that applies dictionary learning to regularize and reconstruct the normal vectors from the images under the classic Lambertian reflectance model.
no code implementations • 30 Sep 2017 • Andrew J. Wagenmaker, Brian E. Moore, Raj Rao Nadakuditi
This paper introduces a novel approach to robust surface reconstruction from photometric stereo normal vector maps that is particularly well-suited for reconstructing surfaces from noisy gradients.
no code implementations • 30 Sep 2017 • Andrew J. Wagenmaker, Brian E. Moore, Raj Rao Nadakuditi
Photometric stereo is a method for estimating the normal vectors of an object from images of the object under varying lighting conditions.
no code implementations • 27 Sep 2017 • Chen Gao, Brian E. Moore, Raj Rao Nadakuditi
This work presents a novel approach for robust PCA with total variation regularization for foreground-background separation and denoising on noisy, moving camera video.
no code implementations • 13 Nov 2016 • Saiprasad Ravishankar, Brian E. Moore, Raj Rao Nadakuditi, Jeffrey A. Fessler
For example, the patches of the underlying data are modeled as sparse in an adaptive dictionary domain, and the resulting image and dictionary estimation from undersampled measurements is called dictionary-blind compressed sensing, or the dynamic image sequence is modeled as a sum of low-rank and sparse (in some transform domain) components (L+S model) that are estimated from limited measurements.
no code implementations • 27 Nov 2015 • Saiprasad Ravishankar, Raj Rao Nadakuditi, Jeffrey A. Fessler
The proposed block coordinate descent algorithm involves efficient closed-form solutions.
1 code implementation • 19 Nov 2015 • Saiprasad Ravishankar, Raj Rao Nadakuditi, Jeffrey A. Fessler
This paper exploits the ideas that drive algorithms such as K-SVD, and investigates in detail efficient methods for aggregate sparsity penalized dictionary learning by first approximating the data with a sum of sparse rank-one matrices (outer products) and then using a block coordinate descent approach to estimate the unknowns.
no code implementations • 25 Jun 2013 • Raj Rao Nadakuditi
The truncated singular value decomposition (SVD) of the measurement matrix is the optimal solution to the_representation_ problem of how to best approximate a noisy measurement matrix using a low-rank matrix.