Skip to main content
OSTI.GOVU.S. Department of Energy Office of Scientific and Technical Information
U.S. Department of Energy
Office of Scientific and Technical Information
 

First- and Second-Order Methods for Online Convolutional Dictionary Learning

Journal Article · · SIAM Journal on Imaging Sciences
DOI:https://doi.org/10.1137/17M1145689· OSTI ID:1475355
 [1];  [2];  [2];  [1]
  1. Univ. of California, Los Angeles, CA (United States). Dept. of Mathematics
  2. Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
+ Show Author Affiliations
Convolutional sparse representations are a form of sparse representation with a structured, translation-invariant dictionary. Most convolutional dictionary learning algorithms to date operate in batch mode, requiring simultaneous access to all training images during the learning process, which results in very high memory usage and severely limits the training data size that can be used. Very recently, however, a number of authors have considered the design of online convolutional dictionary learning algorithms that offer far better scaling of memory and computational cost with training set size than batch methods. This study extends our prior work, improving a number of aspects of our previous algorithm; proposing an entirely new one, with better performance, that supports the inclusion of a spatial mask for learning from incomplete data; and providing a rigorous theoretical analysis of these methods.
Research Organization:
Los Alamos National Laboratory (LANL)
Sponsoring Organization:
LANL Laboratory Directed Research and Development (LDRD) Program; National Science Foundation (NSF) (United States); Office of Naval Research (ONR) (United States); USDOE
Grant/Contract Number:
AC52-06NA25396
OSTI ID:
1475355
Report Number(s):
LA-UR-17-27611
Journal Information:
SIAM Journal on Imaging Sciences, Journal Name: SIAM Journal on Imaging Sciences Journal Issue: 2 Vol. 11; ISSN 1936-4954
Publisher:
Society for Industrial and Applied MathematicsCopyright Statement
Country of Publication:
United States
Language:
English

Figures / Tables (23)

Algorithm 1(p. 9)figure Algorithm 1
Table 1(p. 9)table Table 1
Fig. 1(p. 12)figure Fig. 1
Fig. 2(p. 12)figure Fig. 2
Algorithm 2(p. 14)figure Algorithm 2
Table 2(p. 14)table Table 2
Fig. 3(p. 18)figure Fig. 3
Table 3(p. 18)table Table 3
Fig. 4(p. 19)figure Fig. 4
Fig. 5(p. 19)figure Fig. 5
Fig. 6(p. 20)figure Fig. 6
Fig. 7(p. 20)figure Fig. 7
Fig. 8(p. 21)figure Fig. 8
Fig. 9(p. 22)figure Fig. 9
Table 4(p. 22)table Table 4
Fig. 10(p. 23)figure Fig. 10
Fig. 11(p. 23)figure Fig. 11
Table 5(p. 24)table Table 5
Fig. 12(p. 25)figure Fig. 12
Fig. 13(p. 26)figure Fig. 13
Fig. 14(p. 27)figure Fig. 14
Fig. 15(p. 28)figure Fig. 15
Algorithm 3(p. 30)figure Algorithm 3

Similar Records

Convolutional Dictionary Learning: A Comparative Review and New Algorithms
Journal Article · Fri Aug 31 20:00:00 EDT 2018 · IEEE Transactions on Computational Imaging · OSTI ID:1711366
Convolutional Dictionary Regularizers for Tomographic Inversion
Conference · Wed May 01 00:00:00 EDT 2019 · OSTI ID:1511929
Dictionary Learning with Accumulator Neurons
Conference · Mon Aug 01 00:00:00 EDT 2022 · OSTI ID:1893845

Related Subjects

97 MATHEMATICS AND COMPUTING
Information Science
Mathematics