Matches in SemOpenAlex for { <https://semopenalex.org/work/W3113205501> ?p ?o ?g. }
- W3113205501 abstract "Features obtained from object recognition CNNs have been widely used for measuring perceptual similarities between images. Such differentiable metrics can be used as perceptual learning losses to train image enhancement models. However, the choice of the distance function between input and target features may have a consequential impact on the performance of the trained model. While using the norm of the difference between extracted features leads to limited hallucination of details, measuring the distance between distributions of features may generate more textures; yet also more unrealistic details and artifacts. In this paper, we demonstrate that aggregating 1D-Wasserstein distances between CNN activations is more reliable than the existing approaches, and it can significantly improve the perceptual performance of enhancement models. More explicitly, we show that in imaging applications such as denoising, super-resolution, demosaicing, deblurring and JPEG artifact removal, the proposed learning loss outperforms the current state-of-the-art on reference-based perceptual losses. This means that the proposed learning loss can be plugged into different imaging frameworks and produce perceptually realistic results." @default.
- W3113205501 created "2020-12-21" @default.
- W3113205501 creator A5002085979 @default.
- W3113205501 creator A5008529493 @default.
- W3113205501 creator A5084566603 @default.
- W3113205501 date "2020-12-16" @default.
- W3113205501 modified "2023-09-27" @default.
- W3113205501 title "Projected Distribution Loss for Image Enhancement" @default.
- W3113205501 cites W1580389772 @default.
- W3113205501 cites W1639961155 @default.
- W3113205501 cites W1686810756 @default.
- W3113205501 cites W1885185971 @default.
- W3113205501 cites W1901129140 @default.
- W3113205501 cites W2019106840 @default.
- W3113205501 cites W2099471712 @default.
- W3113205501 cites W2158131535 @default.
- W3113205501 cites W2194775991 @default.
- W3113205501 cites W2331128040 @default.
- W3113205501 cites W2475287302 @default.
- W3113205501 cites W2560533888 @default.
- W3113205501 cites W2562637781 @default.
- W3113205501 cites W2583638424 @default.
- W3113205501 cites W2593729559 @default.
- W3113205501 cites W2741137940 @default.
- W3113205501 cites W2799265886 @default.
- W3113205501 cites W2943960148 @default.
- W3113205501 cites W2961218591 @default.
- W3113205501 cites W2962772087 @default.
- W3113205501 cites W2962785568 @default.
- W3113205501 cites W2963073614 @default.
- W3113205501 cites W2963289467 @default.
- W3113205501 cites W2963312584 @default.
- W3113205501 cites W2963372104 @default.
- W3113205501 cites W2963373786 @default.
- W3113205501 cites W2963398989 @default.
- W3113205501 cites W2963470893 @default.
- W3113205501 cites W2963563295 @default.
- W3113205501 cites W2963869863 @default.
- W3113205501 cites W2964030969 @default.
- W3113205501 cites W2964046669 @default.
- W3113205501 cites W2964168764 @default.
- W3113205501 cites W2964309429 @default.
- W3113205501 cites W2970112944 @default.
- W3113205501 cites W2979557588 @default.
- W3113205501 cites W2982795046 @default.
- W3113205501 cites W3012255272 @default.
- W3113205501 cites W3039256091 @default.
- W3113205501 cites W3103635814 @default.
- W3113205501 cites W3106756503 @default.
- W3113205501 cites W385466589 @default.
- W3113205501 cites W562660536 @default.
- W3113205501 hasPublicationYear "2020" @default.
- W3113205501 type Work @default.
- W3113205501 sameAs 3113205501 @default.
- W3113205501 citedByCount "2" @default.
- W3113205501 countsByYear W31132055012021 @default.
- W3113205501 crossrefType "posted-content" @default.
- W3113205501 hasAuthorship W3113205501A5002085979 @default.
- W3113205501 hasAuthorship W3113205501A5008529493 @default.
- W3113205501 hasAuthorship W3113205501A5084566603 @default.
- W3113205501 hasConcept C106430172 @default.
- W3113205501 hasConcept C115961682 @default.
- W3113205501 hasConcept C134306372 @default.
- W3113205501 hasConcept C138885662 @default.
- W3113205501 hasConcept C153180895 @default.
- W3113205501 hasConcept C154945302 @default.
- W3113205501 hasConcept C163294075 @default.
- W3113205501 hasConcept C169760540 @default.
- W3113205501 hasConcept C202615002 @default.
- W3113205501 hasConcept C26760741 @default.
- W3113205501 hasConcept C2776401178 @default.
- W3113205501 hasConcept C2777693668 @default.
- W3113205501 hasConcept C31972630 @default.
- W3113205501 hasConcept C33923547 @default.
- W3113205501 hasConcept C41008148 @default.
- W3113205501 hasConcept C41895202 @default.
- W3113205501 hasConcept C86803240 @default.
- W3113205501 hasConcept C9417928 @default.
- W3113205501 hasConceptScore W3113205501C106430172 @default.
- W3113205501 hasConceptScore W3113205501C115961682 @default.
- W3113205501 hasConceptScore W3113205501C134306372 @default.
- W3113205501 hasConceptScore W3113205501C138885662 @default.
- W3113205501 hasConceptScore W3113205501C153180895 @default.
- W3113205501 hasConceptScore W3113205501C154945302 @default.
- W3113205501 hasConceptScore W3113205501C163294075 @default.
- W3113205501 hasConceptScore W3113205501C169760540 @default.
- W3113205501 hasConceptScore W3113205501C202615002 @default.
- W3113205501 hasConceptScore W3113205501C26760741 @default.
- W3113205501 hasConceptScore W3113205501C2776401178 @default.
- W3113205501 hasConceptScore W3113205501C2777693668 @default.
- W3113205501 hasConceptScore W3113205501C31972630 @default.
- W3113205501 hasConceptScore W3113205501C33923547 @default.
- W3113205501 hasConceptScore W3113205501C41008148 @default.
- W3113205501 hasConceptScore W3113205501C41895202 @default.
- W3113205501 hasConceptScore W3113205501C86803240 @default.
- W3113205501 hasConceptScore W3113205501C9417928 @default.
- W3113205501 hasLocation W31132055011 @default.
- W3113205501 hasOpenAccess W3113205501 @default.
- W3113205501 hasPrimaryLocation W31132055011 @default.
- W3113205501 hasRelatedWork W2010958786 @default.