Matches in SemOpenAlex for { <https://semopenalex.org/work/W3034822012> ?p ?o ?g. }
Showing items 1 to 70 of
70
with 100 items per page.
- W3034822012 abstract "Reconstruction tasks in computer vision aim fundamentally to recover an undetermined signal from a set of noisy measurements. Examples include super-resolution, image denoising, and non-rigid structure from motion, all of which have seen recent advancements through deep learning. However, earlier work made extensive use of sparse signal reconstruction frameworks (e.g. convolutional sparse coding). While this work was ultimately surpassed by deep learning, it rested on a much more developed theoretical framework. Recent work by Papyan et. al. provides a bridge between the two approaches by showing how a convolutional neural network (CNN) can be viewed as an approximate solution to a convolutional sparse coding (CSC) problem. In this work we argue that for some types of inverse problems the CNN approximation breaks down leading to poor performance. We argue that for these types of problems the CSC approach should be used instead and validate this argument with empirical evidence. Specifically we identify JPEG artifact reduction and non-rigid trajectory reconstruction as challenging inverse problems for CNNs and demonstrate state of the art performance on them using a CSC method. Furthermore, we offer some practical improvements to this model and its application, and also show how insights from the CSC model can be used to make CNNs effective in tasks where their naive application fails." @default.
- W3034822012 created "2020-06-19" @default.
- W3034822012 creator A5007271053 @default.
- W3034822012 creator A5055053078 @default.
- W3034822012 date "2020-06-01" @default.
- W3034822012 modified "2023-10-16" @default.
- W3034822012 title "When to Use Convolutional Neural Networks for Inverse Problems" @default.
- W3034822012 cites W1510019349 @default.
- W3034822012 cites W1885185971 @default.
- W3034822012 cites W1968154520 @default.
- W3034822012 cites W1980119385 @default.
- W3034822012 cites W2075230492 @default.
- W3034822012 cites W2115706991 @default.
- W3034822012 cites W2135046866 @default.
- W3034822012 cites W2153663612 @default.
- W3034822012 cites W2160547390 @default.
- W3034822012 cites W2207282238 @default.
- W3034822012 cites W2604885021 @default.
- W3034822012 cites W2752693045 @default.
- W3034822012 cites W2963047604 @default.
- W3034822012 cites W2963316641 @default.
- W3034822012 cites W2963747696 @default.
- W3034822012 cites W2964268434 @default.
- W3034822012 cites W2982191424 @default.
- W3034822012 cites W2987869089 @default.
- W3034822012 cites W3104720471 @default.
- W3034822012 cites W4206310440 @default.
- W3034822012 cites W4235713725 @default.
- W3034822012 doi "https://doi.org/10.1109/cvpr42600.2020.00825" @default.
- W3034822012 hasPublicationYear "2020" @default.
- W3034822012 type Work @default.
- W3034822012 sameAs 3034822012 @default.
- W3034822012 citedByCount "8" @default.
- W3034822012 countsByYear W30348220122020 @default.
- W3034822012 countsByYear W30348220122021 @default.
- W3034822012 countsByYear W30348220122022 @default.
- W3034822012 crossrefType "proceedings-article" @default.
- W3034822012 hasAuthorship W3034822012A5007271053 @default.
- W3034822012 hasAuthorship W3034822012A5055053078 @default.
- W3034822012 hasBestOaLocation W30348220122 @default.
- W3034822012 hasConcept C154945302 @default.
- W3034822012 hasConcept C207467116 @default.
- W3034822012 hasConcept C2524010 @default.
- W3034822012 hasConcept C33923547 @default.
- W3034822012 hasConcept C41008148 @default.
- W3034822012 hasConcept C81363708 @default.
- W3034822012 hasConceptScore W3034822012C154945302 @default.
- W3034822012 hasConceptScore W3034822012C207467116 @default.
- W3034822012 hasConceptScore W3034822012C2524010 @default.
- W3034822012 hasConceptScore W3034822012C33923547 @default.
- W3034822012 hasConceptScore W3034822012C41008148 @default.
- W3034822012 hasConceptScore W3034822012C81363708 @default.
- W3034822012 hasLocation W30348220121 @default.
- W3034822012 hasLocation W30348220122 @default.
- W3034822012 hasOpenAccess W3034822012 @default.
- W3034822012 hasPrimaryLocation W30348220121 @default.
- W3034822012 hasRelatedWork W2285788670 @default.
- W3034822012 hasRelatedWork W2521062615 @default.
- W3034822012 hasRelatedWork W2735477435 @default.
- W3034822012 hasRelatedWork W2946452775 @default.
- W3034822012 hasRelatedWork W2955938200 @default.
- W3034822012 hasRelatedWork W3016958897 @default.
- W3034822012 hasRelatedWork W3090822330 @default.
- W3034822012 hasRelatedWork W3113517923 @default.
- W3034822012 hasRelatedWork W3181746755 @default.
- W3034822012 hasRelatedWork W4239686595 @default.
- W3034822012 isParatext "false" @default.
- W3034822012 isRetracted "false" @default.
- W3034822012 magId "3034822012" @default.
- W3034822012 workType "article" @default.