Sparse Modeling for Image and Vision Processing

By Julien Mairal, Francis R. Bach, Jean Ponce. Foundations and Trends in Computer Graphics and Vision, 2014

506 citations recorded by the source corpus.

In recent years, a large amount of multi-disciplinary research has been conducted on sparse models and their applications. In statistics and machine learning, the sparsity principle is used to perform model selection - that is, automatically selecting a simple model among a large collection of them. In signal processing, sparse coding consists of representing data with linear combinations of a few dictionary elements. Subsequently, the corresponding tools have been widely adopted by several scientific communities such as neuroscience, bioinformatics, or computer vision. The goal of this monograph is to offer a self-contained view of sparse modeling for visual recognition and image processing. More specifically, we focus on applications where the dictionary is learned and adapted to data, yielding a compact representation that has been successful in various contexts.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/a10a39d2811a804fcead1667c54c4465c3314484

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Efficient Algorithms for Convolutional Sparse Representations (2016, methodology): “[1], [2] is a widely used technique for a very broad range of signal and image processing applications.”
Deep learning ensembles for melanoma recognition in dermoscopy images (2016, methodology): “The SPArse Modeling Software (SPAMS) sparse coding dictionary learning algorithm [44] is an online optimization approach for this method, based on stochastic approximations.”
Machine learning in acoustics: Theory and applications. (2019, methodology): “Clustering-based dictionary learning methods 25 are based on the alternating optimization concept introduced in K-means and EM.”
Solving inverse problems using data-driven models (2019, background): “Mallat (2012), ‘Group invariant scattering’, Commun.”
A Survey on ensemble learning under the era of deep learning (2021, background): “2010) and a type of sparse coding (SC) (Mairal 2014).”
Denoising Gravitational Waves with Enhanced Deep Recurrent Denoising Auto-encoders (2017, methodology): “We show the results of denoising with our model using signals from the test set injected into real LIGO noise in FIG 2, and compare them with PCA and dictionary learning meth-ods (using the code based on [50]).”
Complete dictionary recovery over the sphere (2015, background): “Its applications span classical image processing, visual recognition, compressive signal acquisition, as well as recent deep architectures for signal classification [2, 3].”
Scaling and evaluating sparse autoencoders (2024, unknown): “This approach has been particularly influential in image processing, as seen for example in [Mairal et al., 2014].”
Inducing wavelets into random fields via generative boosting (2016, unknown): “Motivation 1: high-level representations of sparse-land Wavelet sparse coding underlies many successful methods in signal processing, such as image compression, de-noising, and super-resolution [7] [14].”
Machine Learning in Seismology: Turning Data into Insights (2018, methodology): “Methods that do not require ground truths are based on adaptive, unsupervised learning frameworks (Elad, 2010; Mairal et al., 2014).”
Fast Robust PCA on Graphs (2015, unknown): “Fast Robust PCA on Graphs Nauman Shahid ∗ , Nathanael Perraudin, Vassilis Kalofolias, Gilles Puy † , Pierre Vandergheynst Email: { nauman.shahid, Signal Processing nathanael.perraudin, vassilis.kalofolias, pierre.vandergheynst @epﬂ.ch, † gilles.puy@inria.fr Switzerland Laboratory 2”
Convolutional Neural Networks Analyzed via Convolutional Sparse Coding (2016, methodology): “…this model has been commonly used for modeling local patches extracted from the global data mainly due to the computational difficulties related to the task of learning the dictionary (Elad and Aharon, 2006; Dong et al., 2011; Mairal et al., 2014; Romano and Elad, 2015; Sulam and Elad, 2015).”