Matches in SemOpenAlex for { <https://semopenalex.org/work/W4297330762> ?p ?o ?g. }
Showing items 1 to 84 of
84
with 100 items per page.
- W4297330762 abstract "Full-Reference Image Quality Assessment (FR-IQA) metrics such as PSNR, SSIM, and LPIPS have been widely adopted for evaluating image restoration (IR) methods. However, pristine-quality images are usually not available, making inferior No-Reference Image Quality Assessment (NR-IQA) metrics seem to be the only solutions in practical applications. Fortunately, when evaluating image restoration methods, paired degraded and restoration images are generally available. Thus, this paper takes a step forward to develop a Degraded-Reference IQA (DR-IQA) model while respecting its correspondence with FR-IQA metrics. To this end, we adopt a simple encoder-decoder as DR-IQA model, and take paired degraded and restoration images as the input to predict distortion maps guided by FR-IQA metrics. More importantly, due to the diversity and continuous development of image restoration models, it is difficult to make the DR-IQA model learned based on a specific restoration model generalize well to other ones. To address this issue, we augment the DR-IQA training samples by adding the results produced by in-training restoration models. Benefiting from the diversity of training samples, our learned DR-IQA model generalizes well to unseen restoration models. We respectively test our DR-IQA models on various image restoration tasks,e.g., denoising, super-resolution, JPEG deblocking, and complicated degradations, where our method can further close the performance gap between FR-IQA metrics and the state-of-the-art NR-IQA methods. Moreover, experiments also show the effectiveness of our method in performance comparison and model selection of image restoration models without ground-truth clean images. Source code will be made publicly available." @default.
- W4297330762 created "2022-09-28" @default.
- W4297330762 creator A5018318136 @default.
- W4297330762 creator A5028650738 @default.
- W4297330762 creator A5051402211 @default.
- W4297330762 creator A5075302729 @default.
- W4297330762 date "2022-10-10" @default.
- W4297330762 modified "2023-10-01" @default.
- W4297330762 title "In-training Restoration Models Matter" @default.
- W4297330762 cites W1977725648 @default.
- W4297330762 cites W1982471090 @default.
- W4297330762 cites W2046119925 @default.
- W4297330762 cites W2056370875 @default.
- W4297330762 cites W2129644086 @default.
- W4297330762 cites W2130184048 @default.
- W4297330762 cites W2133665775 @default.
- W4297330762 cites W2142884912 @default.
- W4297330762 cites W2161907179 @default.
- W4297330762 cites W2162915697 @default.
- W4297330762 cites W2473697052 @default.
- W4297330762 cites W2739757502 @default.
- W4297330762 cites W2962785568 @default.
- W4297330762 cites W2963470893 @default.
- W4297330762 cites W2963494934 @default.
- W4297330762 cites W2963796134 @default.
- W4297330762 cites W2964065910 @default.
- W4297330762 cites W3034882073 @default.
- W4297330762 cites W3035719652 @default.
- W4297330762 cites W3098560717 @default.
- W4297330762 cites W3103635814 @default.
- W4297330762 cites W3104725225 @default.
- W4297330762 cites W3171175128 @default.
- W4297330762 cites W3175517129 @default.
- W4297330762 cites W3176592254 @default.
- W4297330762 cites W4242059867 @default.
- W4297330762 doi "https://doi.org/10.1145/3552456.3555667" @default.
- W4297330762 hasPublicationYear "2022" @default.
- W4297330762 type Work @default.
- W4297330762 citedByCount "0" @default.
- W4297330762 crossrefType "proceedings-article" @default.
- W4297330762 hasAuthorship W4297330762A5018318136 @default.
- W4297330762 hasAuthorship W4297330762A5028650738 @default.
- W4297330762 hasAuthorship W4297330762A5051402211 @default.
- W4297330762 hasAuthorship W4297330762A5075302729 @default.
- W4297330762 hasConcept C106430172 @default.
- W4297330762 hasConcept C115961682 @default.
- W4297330762 hasConcept C126780896 @default.
- W4297330762 hasConcept C153180895 @default.
- W4297330762 hasConcept C154945302 @default.
- W4297330762 hasConcept C194257627 @default.
- W4297330762 hasConcept C2776257435 @default.
- W4297330762 hasConcept C31258907 @default.
- W4297330762 hasConcept C31972630 @default.
- W4297330762 hasConcept C41008148 @default.
- W4297330762 hasConcept C55020928 @default.
- W4297330762 hasConcept C9417928 @default.
- W4297330762 hasConceptScore W4297330762C106430172 @default.
- W4297330762 hasConceptScore W4297330762C115961682 @default.
- W4297330762 hasConceptScore W4297330762C126780896 @default.
- W4297330762 hasConceptScore W4297330762C153180895 @default.
- W4297330762 hasConceptScore W4297330762C154945302 @default.
- W4297330762 hasConceptScore W4297330762C194257627 @default.
- W4297330762 hasConceptScore W4297330762C2776257435 @default.
- W4297330762 hasConceptScore W4297330762C31258907 @default.
- W4297330762 hasConceptScore W4297330762C31972630 @default.
- W4297330762 hasConceptScore W4297330762C41008148 @default.
- W4297330762 hasConceptScore W4297330762C55020928 @default.
- W4297330762 hasConceptScore W4297330762C9417928 @default.
- W4297330762 hasLocation W42973307621 @default.
- W4297330762 hasOpenAccess W4297330762 @default.
- W4297330762 hasPrimaryLocation W42973307621 @default.
- W4297330762 hasRelatedWork W2130228941 @default.
- W4297330762 hasRelatedWork W2132132164 @default.
- W4297330762 hasRelatedWork W2370063699 @default.
- W4297330762 hasRelatedWork W2574052219 @default.
- W4297330762 hasRelatedWork W2906703443 @default.
- W4297330762 hasRelatedWork W3001963043 @default.
- W4297330762 hasRelatedWork W3044651058 @default.
- W4297330762 hasRelatedWork W3152929143 @default.
- W4297330762 hasRelatedWork W3217021205 @default.
- W4297330762 hasRelatedWork W4236771684 @default.
- W4297330762 isParatext "false" @default.
- W4297330762 isRetracted "false" @default.
- W4297330762 workType "article" @default.