Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100456687> ?p ?o ?g. }
- W3100456687 endingPage "373" @default.
- W3100456687 startingPage "360" @default.
- W3100456687 abstract "The use of ℓ <sub xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>p</sub> (p = 1,2) norms has largely dominated the measurement of loss in neural networks due to their simplicity and analytical properties. However, when used to assess the loss of visual information, these simple norms are not very consistent with human perception. Here, we describe a different “proximal” approach to optimize image analysis networks against quantitative perceptual models. Specifically, we construct a proxy network, broadly termed ProxIQA, which mimics the perceptual model while serving as a loss layer of the network. We experimentally demonstrate how this optimization framework can be applied to train an end-to-end optimized image compression network. By building on top of an existing deep image compression model, we are able to demonstrate a bitrate reduction of as much as 31% over MSE optimization, given a specified perceptual quality (VMAF) level." @default.
- W3100456687 created "2020-11-23" @default.
- W3100456687 creator A5009552579 @default.
- W3100456687 creator A5036706118 @default.
- W3100456687 creator A5040009629 @default.
- W3100456687 creator A5075463806 @default.
- W3100456687 creator A5056447210 @default.
- W3100456687 date "2021-01-01" @default.
- W3100456687 modified "2023-09-29" @default.
- W3100456687 title "ProxIQA: A Proxy Approach to Perceptual Optimization of Learned Image Compression" @default.
- W3100456687 cites W1580389772 @default.
- W3100456687 cites W1607233498 @default.
- W3100456687 cites W2009272644 @default.
- W3100456687 cites W2013784666 @default.
- W3100456687 cites W2015196405 @default.
- W3100456687 cites W2037642501 @default.
- W3100456687 cites W2046119925 @default.
- W3100456687 cites W2099474155 @default.
- W3100456687 cites W2103997799 @default.
- W3100456687 cites W2108598243 @default.
- W3100456687 cites W2129768577 @default.
- W3100456687 cites W2133665775 @default.
- W3100456687 cites W2134227958 @default.
- W3100456687 cites W2137136191 @default.
- W3100456687 cites W2141983208 @default.
- W3100456687 cites W2164287382 @default.
- W3100456687 cites W2178576682 @default.
- W3100456687 cites W2194775991 @default.
- W3100456687 cites W2395611524 @default.
- W3100456687 cites W2550673061 @default.
- W3100456687 cites W2556068545 @default.
- W3100456687 cites W2557414982 @default.
- W3100456687 cites W2562637781 @default.
- W3100456687 cites W2586275201 @default.
- W3100456687 cites W2597747080 @default.
- W3100456687 cites W2607041014 @default.
- W3100456687 cites W2768814045 @default.
- W3100456687 cites W2904829482 @default.
- W3100456687 cites W2905530235 @default.
- W3100456687 cites W2923664668 @default.
- W3100456687 cites W2948654165 @default.
- W3100456687 cites W2955863859 @default.
- W3100456687 cites W2962785568 @default.
- W3100456687 cites W2963037581 @default.
- W3100456687 cites W2963149687 @default.
- W3100456687 cites W2963189365 @default.
- W3100456687 cites W2963470893 @default.
- W3100456687 cites W2963711615 @default.
- W3100456687 cites W2963713691 @default.
- W3100456687 cites W2973013219 @default.
- W3100456687 cites W2981613960 @default.
- W3100456687 cites W2983979703 @default.
- W3100456687 cites W3015735582 @default.
- W3100456687 cites W3015863973 @default.
- W3100456687 cites W3088230179 @default.
- W3100456687 cites W3090186338 @default.
- W3100456687 cites W3100498948 @default.
- W3100456687 cites W3103682398 @default.
- W3100456687 cites W764651262 @default.
- W3100456687 doi "https://doi.org/10.1109/tip.2020.3036752" @default.
- W3100456687 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/33186110" @default.
- W3100456687 hasPublicationYear "2021" @default.
- W3100456687 type Work @default.
- W3100456687 sameAs 3100456687 @default.
- W3100456687 citedByCount "36" @default.
- W3100456687 countsByYear W31004566872020 @default.
- W3100456687 countsByYear W31004566872021 @default.
- W3100456687 countsByYear W31004566872022 @default.
- W3100456687 countsByYear W31004566872023 @default.
- W3100456687 crossrefType "journal-article" @default.
- W3100456687 hasAuthorship W3100456687A5009552579 @default.
- W3100456687 hasAuthorship W3100456687A5036706118 @default.
- W3100456687 hasAuthorship W3100456687A5040009629 @default.
- W3100456687 hasAuthorship W3100456687A5056447210 @default.
- W3100456687 hasAuthorship W3100456687A5075463806 @default.
- W3100456687 hasBestOaLocation W31004566872 @default.
- W3100456687 hasConcept C115961682 @default.
- W3100456687 hasConcept C119857082 @default.
- W3100456687 hasConcept C124101348 @default.
- W3100456687 hasConcept C13481523 @default.
- W3100456687 hasConcept C153180895 @default.
- W3100456687 hasConcept C154945302 @default.
- W3100456687 hasConcept C169760540 @default.
- W3100456687 hasConcept C26760741 @default.
- W3100456687 hasConcept C2780148112 @default.
- W3100456687 hasConcept C41008148 @default.
- W3100456687 hasConcept C50644808 @default.
- W3100456687 hasConcept C55020928 @default.
- W3100456687 hasConcept C78548338 @default.
- W3100456687 hasConcept C86803240 @default.
- W3100456687 hasConcept C9417928 @default.
- W3100456687 hasConceptScore W3100456687C115961682 @default.
- W3100456687 hasConceptScore W3100456687C119857082 @default.
- W3100456687 hasConceptScore W3100456687C124101348 @default.
- W3100456687 hasConceptScore W3100456687C13481523 @default.
- W3100456687 hasConceptScore W3100456687C153180895 @default.
- W3100456687 hasConceptScore W3100456687C154945302 @default.
- W3100456687 hasConceptScore W3100456687C169760540 @default.