Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100506797> ?p ?o ?g. }
- W3100506797 endingPage "103134" @default.
- W3100506797 startingPage "103134" @default.
- W3100506797 abstract "Neural networks (NNs) are becoming the tool of choice for sharpening blurred images. We discuss and categorize deblurring NNs. Then we evaluate seven NNs for non-blind deblurring (NBD), and seven NNs and four optimization techniques for blind deblurring (BD). To do this we use several current datasets containing pairs of sharp and blurred images, synthesized either by convolving sharp images with blur kernels or by averaging consecutive sharp images, so as to produce both uniform and non-uniform blurs. We also introduce a newly reorganized benchmark dataset in which blurred images have been classified using attributes that depend on the extent of the blur. We use this dataset to compare the effectiveness of single and multi-scale training in coping with large blurs. On NBD, NNs that use regularization with a denoising prior network outperform other denoising NNs; and NNs that use a deep image prior network outperform other deconvolution NNs. On BD, NNs outperform optimizations in signal-difference terms, but not in terms of perceptual fidelity. We found that multi-scale training helps NNs to deal with large blurs, and RNNs outperform CNNs. We also observed that GANs using a perceptual loss function produce artifacts; but also that some form of perceptual fidelity loss is required to get the best results from NNs. We contend that the domain bias of current datasets works against robustness and generality. And we discuss the potential of more sophisticated perceptual loss functions, attention techniques, and unsupervised learning." @default.
- W3100506797 created "2020-11-23" @default.
- W3100506797 creator A5009485352 @default.
- W3100506797 creator A5076640352 @default.
- W3100506797 creator A5086877012 @default.
- W3100506797 date "2021-02-01" @default.
- W3100506797 modified "2023-10-18" @default.
- W3100506797 title "Single-image deblurring with neural networks: A comparative survey" @default.
- W3100506797 cites W1457323852 @default.
- W3100506797 cites W1510355813 @default.
- W3100506797 cites W1916935112 @default.
- W3100506797 cites W1973567017 @default.
- W3100506797 cites W1980884128 @default.
- W3100506797 cites W1996726072 @default.
- W3100506797 cites W2009501432 @default.
- W3100506797 cites W2010771616 @default.
- W3100506797 cites W2043529138 @default.
- W3100506797 cites W2044945560 @default.
- W3100506797 cites W2047123483 @default.
- W3100506797 cites W2056370875 @default.
- W3100506797 cites W2077931794 @default.
- W3100506797 cites W2088909704 @default.
- W3100506797 cites W2094881235 @default.
- W3100506797 cites W2114122776 @default.
- W3100506797 cites W2121689659 @default.
- W3100506797 cites W2126252886 @default.
- W3100506797 cites W2133665775 @default.
- W3100506797 cites W2133957619 @default.
- W3100506797 cites W2138141351 @default.
- W3100506797 cites W2141115311 @default.
- W3100506797 cites W2150060382 @default.
- W3100506797 cites W2163895617 @default.
- W3100506797 cites W2166576327 @default.
- W3100506797 cites W2167307343 @default.
- W3100506797 cites W2170608748 @default.
- W3100506797 cites W2194775991 @default.
- W3100506797 cites W2242218935 @default.
- W3100506797 cites W2267317359 @default.
- W3100506797 cites W2319561215 @default.
- W3100506797 cites W2331376995 @default.
- W3100506797 cites W2461759225 @default.
- W3100506797 cites W2465552163 @default.
- W3100506797 cites W2474628748 @default.
- W3100506797 cites W2552111036 @default.
- W3100506797 cites W2560533888 @default.
- W3100506797 cites W2564023417 @default.
- W3100506797 cites W2573726823 @default.
- W3100506797 cites W2601564443 @default.
- W3100506797 cites W2611015177 @default.
- W3100506797 cites W2613155248 @default.
- W3100506797 cites W2738579427 @default.
- W3100506797 cites W2740543610 @default.
- W3100506797 cites W2768610172 @default.
- W3100506797 cites W2798581339 @default.
- W3100506797 cites W2798735168 @default.
- W3100506797 cites W2798840374 @default.
- W3100506797 cites W2895327470 @default.
- W3100506797 cites W2905658590 @default.
- W3100506797 cites W2963091558 @default.
- W3100506797 cites W2963312584 @default.
- W3100506797 cites W2963470893 @default.
- W3100506797 cites W2963667985 @default.
- W3100506797 cites W2963782415 @default.
- W3100506797 cites W2964030969 @default.
- W3100506797 cites W2964040059 @default.
- W3100506797 cites W2964051148 @default.
- W3100506797 cites W2964317599 @default.
- W3100506797 cites W2967205412 @default.
- W3100506797 cites W2985368596 @default.
- W3100506797 cites W3101787898 @default.
- W3100506797 doi "https://doi.org/10.1016/j.cviu.2020.103134" @default.
- W3100506797 hasPublicationYear "2021" @default.
- W3100506797 type Work @default.
- W3100506797 sameAs 3100506797 @default.
- W3100506797 citedByCount "41" @default.
- W3100506797 countsByYear W31005067972021 @default.
- W3100506797 countsByYear W31005067972022 @default.
- W3100506797 countsByYear W31005067972023 @default.
- W3100506797 crossrefType "journal-article" @default.
- W3100506797 hasAuthorship W3100506797A5009485352 @default.
- W3100506797 hasAuthorship W3100506797A5076640352 @default.
- W3100506797 hasAuthorship W3100506797A5086877012 @default.
- W3100506797 hasConcept C104317684 @default.
- W3100506797 hasConcept C106430172 @default.
- W3100506797 hasConcept C11413529 @default.
- W3100506797 hasConcept C115961682 @default.
- W3100506797 hasConcept C119857082 @default.
- W3100506797 hasConcept C153180895 @default.
- W3100506797 hasConcept C154945302 @default.
- W3100506797 hasConcept C174576160 @default.
- W3100506797 hasConcept C185592680 @default.
- W3100506797 hasConcept C2776459999 @default.
- W3100506797 hasConcept C2777693668 @default.
- W3100506797 hasConcept C2781137444 @default.
- W3100506797 hasConcept C31972630 @default.
- W3100506797 hasConcept C41008148 @default.
- W3100506797 hasConcept C50644808 @default.
- W3100506797 hasConcept C55493867 @default.