Matches in SemOpenAlex for { <https://semopenalex.org/work/W3090227645> ?p ?o ?g. }
Showing items 1 to 76 of
76
with 100 items per page.
- W3090227645 abstract "Computer vision datasets containing multiple modalities such as color, depth, and thermal properties are now commonly accessible and useful for solving a wide array of challenging tasks. However, deploying multi-sensor heads is not possible in many scenarios. As such many practical solutions tend to be based on simpler sensors, mostly for cost, simplicity and robustness considerations. In this work, we propose a training methodology to take advantage of these additional modalities available in datasets, even if they are not available at test time. By assuming that the modalities have a strong spatial correlation, we propose Input Dropout, a simple technique that consists in stochastic hiding of one or many input modalities at training time, while using only the canonical (e.g. RGB) modalities at test time. We demonstrate that Input Dropout trivially combines with existing deep convolutional architectures, and improves their performance on a wide range of computer vision tasks such as dehazing, 6-DOF object tracking, pedestrian detection and object classification." @default.
- W3090227645 created "2020-10-08" @default.
- W3090227645 creator A5012606778 @default.
- W3090227645 creator A5034761030 @default.
- W3090227645 creator A5042971282 @default.
- W3090227645 creator A5045218915 @default.
- W3090227645 date "2020-10-01" @default.
- W3090227645 modified "2023-09-27" @default.
- W3090227645 title "Input Dropout for Spatially Aligned Modalities" @default.
- W3090227645 cites W1587447546 @default.
- W3090227645 cites W1910108985 @default.
- W3090227645 cites W1985238052 @default.
- W3090227645 cites W2056952330 @default.
- W3090227645 cites W2067562626 @default.
- W3090227645 cites W2108598243 @default.
- W3090227645 cites W2150066425 @default.
- W3090227645 cites W2194775991 @default.
- W3090227645 cites W2340897893 @default.
- W3090227645 cites W2463402750 @default.
- W3090227645 cites W2565639579 @default.
- W3090227645 cites W2748021867 @default.
- W3090227645 cites W2788682721 @default.
- W3090227645 cites W2798876216 @default.
- W3090227645 cites W2809440904 @default.
- W3090227645 cites W2885826602 @default.
- W3090227645 cites W2897611955 @default.
- W3090227645 cites W2899026108 @default.
- W3090227645 cites W2918587444 @default.
- W3090227645 cites W2963192057 @default.
- W3090227645 cites W2963351448 @default.
- W3090227645 cites W2963504620 @default.
- W3090227645 cites W3100811978 @default.
- W3090227645 cites W4230005465 @default.
- W3090227645 doi "https://doi.org/10.1109/icip40778.2020.9190858" @default.
- W3090227645 hasPublicationYear "2020" @default.
- W3090227645 type Work @default.
- W3090227645 sameAs 3090227645 @default.
- W3090227645 citedByCount "3" @default.
- W3090227645 countsByYear W30902276452020 @default.
- W3090227645 countsByYear W30902276452022 @default.
- W3090227645 crossrefType "proceedings-article" @default.
- W3090227645 hasAuthorship W3090227645A5012606778 @default.
- W3090227645 hasAuthorship W3090227645A5034761030 @default.
- W3090227645 hasAuthorship W3090227645A5042971282 @default.
- W3090227645 hasAuthorship W3090227645A5045218915 @default.
- W3090227645 hasBestOaLocation W30902276452 @default.
- W3090227645 hasConcept C119857082 @default.
- W3090227645 hasConcept C144024400 @default.
- W3090227645 hasConcept C2776145597 @default.
- W3090227645 hasConcept C2779903281 @default.
- W3090227645 hasConcept C36289849 @default.
- W3090227645 hasConcept C41008148 @default.
- W3090227645 hasConceptScore W3090227645C119857082 @default.
- W3090227645 hasConceptScore W3090227645C144024400 @default.
- W3090227645 hasConceptScore W3090227645C2776145597 @default.
- W3090227645 hasConceptScore W3090227645C2779903281 @default.
- W3090227645 hasConceptScore W3090227645C36289849 @default.
- W3090227645 hasConceptScore W3090227645C41008148 @default.
- W3090227645 hasLocation W30902276451 @default.
- W3090227645 hasLocation W30902276452 @default.
- W3090227645 hasOpenAccess W3090227645 @default.
- W3090227645 hasPrimaryLocation W30902276451 @default.
- W3090227645 hasRelatedWork W1968430047 @default.
- W3090227645 hasRelatedWork W2008236508 @default.
- W3090227645 hasRelatedWork W2467325985 @default.
- W3090227645 hasRelatedWork W2888540803 @default.
- W3090227645 hasRelatedWork W3004550220 @default.
- W3090227645 hasRelatedWork W3093343994 @default.
- W3090227645 hasRelatedWork W3134092202 @default.
- W3090227645 hasRelatedWork W3200945463 @default.
- W3090227645 hasRelatedWork W4287869739 @default.
- W3090227645 hasRelatedWork W60093136 @default.
- W3090227645 isParatext "false" @default.
- W3090227645 isRetracted "false" @default.
- W3090227645 magId "3090227645" @default.
- W3090227645 workType "article" @default.