Matches in SemOpenAlex for { <https://semopenalex.org/work/W4303684013> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4303684013 abstract "Over parameterization is a common technique in deep learning to help models learn and generalize sufficiently to the given task; nonetheless, this often leads to enormous network structures and consumes considerable computing resources during training. Recent powerful transformer-based deep learning models on vision tasks usually have heavy parameters and bear training difficulty. However, many dense-prediction low-level computer vision tasks, such as rain streak removing, often need to be executed on devices with limited computing power and memory in practice. Hence, we introduce a recursive local window-based self-attention structure with residual connections and propose deraining a recursive transformer (DRT), which enjoys the superiority of the transformer but requires a small amount of computing resources. In particular, through recursive architecture, our proposed model uses only 1.3% of the number of parameters of the current best performing model in deraining while exceeding the state-of-the-art methods on the Rain100L benchmark by at least 0.33 dB. Ablation studies also investigate the impact of recursions on derain outcomes. Moreover, since the model contains no deliberate design for deraining, it can also be applied to other image restoration tasks. Our experiment shows that it can achieve competitive results on desnowing. The source code and pretrained model can be found at https://github.com/YC-Liang/DRT." @default.
- W4303684013 created "2022-10-09" @default.
- W4303684013 creator A5023363049 @default.
- W4303684013 creator A5024397686 @default.
- W4303684013 creator A5084633048 @default.
- W4303684013 date "2022-04-24" @default.
- W4303684013 modified "2023-09-24" @default.
- W4303684013 title "DRT: A Lightweight Single Image Deraining Recursive Transformer" @default.
- W4303684013 doi "https://doi.org/10.48550/arxiv.2204.11385" @default.
- W4303684013 hasPublicationYear "2022" @default.
- W4303684013 type Work @default.
- W4303684013 citedByCount "0" @default.
- W4303684013 crossrefType "posted-content" @default.
- W4303684013 hasAuthorship W4303684013A5023363049 @default.
- W4303684013 hasAuthorship W4303684013A5024397686 @default.
- W4303684013 hasAuthorship W4303684013A5084633048 @default.
- W4303684013 hasBestOaLocation W43036840131 @default.
- W4303684013 hasConcept C108583219 @default.
- W4303684013 hasConcept C113775141 @default.
- W4303684013 hasConcept C11413529 @default.
- W4303684013 hasConcept C119599485 @default.
- W4303684013 hasConcept C119857082 @default.
- W4303684013 hasConcept C120665830 @default.
- W4303684013 hasConcept C121332964 @default.
- W4303684013 hasConcept C127413603 @default.
- W4303684013 hasConcept C13280743 @default.
- W4303684013 hasConcept C154945302 @default.
- W4303684013 hasConcept C155512373 @default.
- W4303684013 hasConcept C165801399 @default.
- W4303684013 hasConcept C177264268 @default.
- W4303684013 hasConcept C185798385 @default.
- W4303684013 hasConcept C199360897 @default.
- W4303684013 hasConcept C205649164 @default.
- W4303684013 hasConcept C2776760102 @default.
- W4303684013 hasConcept C41008148 @default.
- W4303684013 hasConcept C43126263 @default.
- W4303684013 hasConcept C65185188 @default.
- W4303684013 hasConcept C66322947 @default.
- W4303684013 hasConceptScore W4303684013C108583219 @default.
- W4303684013 hasConceptScore W4303684013C113775141 @default.
- W4303684013 hasConceptScore W4303684013C11413529 @default.
- W4303684013 hasConceptScore W4303684013C119599485 @default.
- W4303684013 hasConceptScore W4303684013C119857082 @default.
- W4303684013 hasConceptScore W4303684013C120665830 @default.
- W4303684013 hasConceptScore W4303684013C121332964 @default.
- W4303684013 hasConceptScore W4303684013C127413603 @default.
- W4303684013 hasConceptScore W4303684013C13280743 @default.
- W4303684013 hasConceptScore W4303684013C154945302 @default.
- W4303684013 hasConceptScore W4303684013C155512373 @default.
- W4303684013 hasConceptScore W4303684013C165801399 @default.
- W4303684013 hasConceptScore W4303684013C177264268 @default.
- W4303684013 hasConceptScore W4303684013C185798385 @default.
- W4303684013 hasConceptScore W4303684013C199360897 @default.
- W4303684013 hasConceptScore W4303684013C205649164 @default.
- W4303684013 hasConceptScore W4303684013C2776760102 @default.
- W4303684013 hasConceptScore W4303684013C41008148 @default.
- W4303684013 hasConceptScore W4303684013C43126263 @default.
- W4303684013 hasConceptScore W4303684013C65185188 @default.
- W4303684013 hasConceptScore W4303684013C66322947 @default.
- W4303684013 hasLocation W43036840131 @default.
- W4303684013 hasOpenAccess W4303684013 @default.
- W4303684013 hasPrimaryLocation W43036840131 @default.
- W4303684013 hasRelatedWork W2088830868 @default.
- W4303684013 hasRelatedWork W2108156218 @default.
- W4303684013 hasRelatedWork W2747898905 @default.
- W4303684013 hasRelatedWork W3026432413 @default.
- W4303684013 hasRelatedWork W4223943233 @default.
- W4303684013 hasRelatedWork W4224922575 @default.
- W4303684013 hasRelatedWork W4303684013 @default.
- W4303684013 hasRelatedWork W4312200629 @default.
- W4303684013 hasRelatedWork W4360585206 @default.
- W4303684013 hasRelatedWork W4364306694 @default.
- W4303684013 isParatext "false" @default.
- W4303684013 isRetracted "false" @default.
- W4303684013 workType "article" @default.