Published on Mon Feb 01 2016

Learning Data Triage: Linear Decoding Works for Compressive MRI

Yen-Huan Li, Volkan Cevher

The standard approach to compressive sampling considers recovering an unknown deterministic signal with certain known structure. We learn a good sub sampling pattern based on available training signals, and reconstruct an accordingly sub-sampled signal. We provide a theoretical guarantee on the recovery error.

0
0
0
Abstract

The standard approach to compressive sampling considers recovering an unknown deterministic signal with certain known structure, and designing the sub-sampling pattern and recovery algorithm based on the known structure. This approach requires looking for a good representation that reveals the signal structure, and solving a non-smooth convex minimization problem (e.g., basis pursuit). In this paper, another approach is considered: We learn a good sub-sampling pattern based on available training signals, without knowing the signal structure in advance, and reconstruct an accordingly sub-sampled signal by computationally much cheaper linear reconstruction. We provide a theoretical guarantee on the recovery error, and show via experiments on real-world MRI data the effectiveness of the proposed compressive MRI scheme.

Mon Jun 04 2012
Machine Learning
Multi-Sparse Signal Recovery for Compressive Sensing
Signal recovery is one of the key techniques of Compressive sensing (CS) Classical methods exploit the sparsity in one domain to formulate the L0 norm optimization. Recent investigation shows that some signals are sparse in multiple domains.
0
0
0
Thu Aug 13 2015
Machine Learning
Optimized Projections for Compressed Sensing via Direct Mutual Coherence Minimization
Compressed Sensing is a novel technique for simultaneous signal sampling and compression. It is based on the existence of a sparse representation of signal and a projected dictionary The dictionary is expected to be of low mutual coherence.
0
0
0
Sun Aug 02 2015
Computer Vision
Dictionary and Image Recovery from Incomplete and Random Measurements
This paper tackles algorithmic and theoretical aspects of dictionary learning from incomplete and random block-wise image measurements. We show that applying a slightly modified dictionary learning algorithm over compressed measurements results in accurate estimation of the ideal dictionary for large-scale images.
0
0
0
Mon Aug 24 2020
Machine Learning
Compressive Phase Retrieval: Optimal Sample Complexity with Deep Generative Priors
Advances in compressive sensing have provided reconstruction algorithms of sparse signals from linear measurements with optimal sample complexity. But extensions of this methodology to nonlinear inverse problems have been met with potentially fundamental sample complexity bottlenecks. In this work, we demonstrate that generative priors can lead to a fundamental advance by permitting optimal sampled complexity by a tractable algorithm.
0
0
0
Wed Apr 30 2014
Computer Vision
Image Compressive Sensing Recovery Using Adaptively Learned Sparsifying Basis via L0 Minimization
Compressive sensing (CS) theory demonstrates that, a signal can be.reconstructed with high probability when it exhibits sparsity in some domain. Most of the conventional CS recovery approaches exploited a set of fixed bases for the entirety of a signal. The intrinsic sparsity of natural images is
0
0
0
Mon Nov 28 2016
Machine Learning
Dictionary Learning with Equiprobable Matching Pursuit
Sparse signal representations based on linear combinations of learned atoms have been used to obtain state-of-the-art results in several practical signal processing applications. Here we study greedy algorithms for unsupervised learning of dictionaries of shift-invariant atoms and propose a new method where each
0
0
0