Matches in SemOpenAlex for { <https://semopenalex.org/work/W3045872183> ?p ?o ?g. }
Showing items 1 to 87 of
87
with 100 items per page.
- W3045872183 abstract "For many small- and medium-vocabulary tasks, audio-visual speech recognition can significantly improve the recognition rates compared to audio-only systems. However, there is still an ongoing debate regarding the best combination strategy for multi-modal information, which should allow for the translation of these gains to large-vocabulary recognition. While an integration at the level of state-posterior probabilities, using dynamic stream weighting, is almost universally helpful for small-vocabulary systems, in large-vocabulary speech recognition, the recognition accuracy remains difficult to improve. In the following, we specifically consider the large-vocabulary task of the LRS2 database, and we investigate a broad range of integration strategies, comparing early integration and end-to-end learning with many versions of hybrid recognition and dynamic stream weighting. One aspect, which is shown to provide much benefit here, is the use of dynamic stream reliability indicators, which allow for hybrid architectures to strongly profit from the inclusion of visual information whenever the audio channel is distorted even slightly." @default.
- W3045872183 created "2020-08-03" @default.
- W3045872183 creator A5007017640 @default.
- W3045872183 creator A5019861809 @default.
- W3045872183 creator A5046724251 @default.
- W3045872183 date "2020-07-28" @default.
- W3045872183 modified "2023-09-27" @default.
- W3045872183 title "Multimodal Integration for Large-Vocabulary Audio-Visual Speech Recognition" @default.
- W3045872183 cites W1524333225 @default.
- W3045872183 cites W2023334087 @default.
- W3045872183 cites W2024490110 @default.
- W3045872183 cites W2098562545 @default.
- W3045872183 cites W2103621378 @default.
- W3045872183 cites W2104263160 @default.
- W3045872183 cites W2113642685 @default.
- W3045872183 cites W2144121180 @default.
- W3045872183 cites W2155998647 @default.
- W3045872183 cites W2219249508 @default.
- W3045872183 cites W2406799669 @default.
- W3045872183 cites W2521686623 @default.
- W3045872183 cites W2586219009 @default.
- W3045872183 cites W2696731410 @default.
- W3045872183 cites W2889448058 @default.
- W3045872183 cites W2889624961 @default.
- W3045872183 cites W2890952074 @default.
- W3045872183 cites W2952746495 @default.
- W3045872183 cites W2963528589 @default.
- W3045872183 cites W2964308564 @default.
- W3045872183 cites W3011234510 @default.
- W3045872183 cites W3103005696 @default.
- W3045872183 doi "https://doi.org/10.48550/arxiv.2007.14223" @default.
- W3045872183 hasPublicationYear "2020" @default.
- W3045872183 type Work @default.
- W3045872183 sameAs 3045872183 @default.
- W3045872183 citedByCount "3" @default.
- W3045872183 countsByYear W30458721832020 @default.
- W3045872183 countsByYear W30458721832021 @default.
- W3045872183 crossrefType "posted-content" @default.
- W3045872183 hasAuthorship W3045872183A5007017640 @default.
- W3045872183 hasAuthorship W3045872183A5019861809 @default.
- W3045872183 hasAuthorship W3045872183A5046724251 @default.
- W3045872183 hasBestOaLocation W30458721831 @default.
- W3045872183 hasConcept C126838900 @default.
- W3045872183 hasConcept C138885662 @default.
- W3045872183 hasConcept C154945302 @default.
- W3045872183 hasConcept C157968479 @default.
- W3045872183 hasConcept C183115368 @default.
- W3045872183 hasConcept C204201278 @default.
- W3045872183 hasConcept C204321447 @default.
- W3045872183 hasConcept C2777601683 @default.
- W3045872183 hasConcept C28490314 @default.
- W3045872183 hasConcept C41008148 @default.
- W3045872183 hasConcept C41895202 @default.
- W3045872183 hasConcept C61328038 @default.
- W3045872183 hasConcept C71924100 @default.
- W3045872183 hasConcept C95623464 @default.
- W3045872183 hasConceptScore W3045872183C126838900 @default.
- W3045872183 hasConceptScore W3045872183C138885662 @default.
- W3045872183 hasConceptScore W3045872183C154945302 @default.
- W3045872183 hasConceptScore W3045872183C157968479 @default.
- W3045872183 hasConceptScore W3045872183C183115368 @default.
- W3045872183 hasConceptScore W3045872183C204201278 @default.
- W3045872183 hasConceptScore W3045872183C204321447 @default.
- W3045872183 hasConceptScore W3045872183C2777601683 @default.
- W3045872183 hasConceptScore W3045872183C28490314 @default.
- W3045872183 hasConceptScore W3045872183C41008148 @default.
- W3045872183 hasConceptScore W3045872183C41895202 @default.
- W3045872183 hasConceptScore W3045872183C61328038 @default.
- W3045872183 hasConceptScore W3045872183C71924100 @default.
- W3045872183 hasConceptScore W3045872183C95623464 @default.
- W3045872183 hasLocation W30458721831 @default.
- W3045872183 hasOpenAccess W3045872183 @default.
- W3045872183 hasPrimaryLocation W30458721831 @default.
- W3045872183 hasRelatedWork W1208717 @default.
- W3045872183 hasRelatedWork W12168553 @default.
- W3045872183 hasRelatedWork W14247858 @default.
- W3045872183 hasRelatedWork W2122633 @default.
- W3045872183 hasRelatedWork W2308727 @default.
- W3045872183 hasRelatedWork W7082836 @default.
- W3045872183 hasRelatedWork W7914383 @default.
- W3045872183 hasRelatedWork W8738421 @default.
- W3045872183 hasRelatedWork W9638664 @default.
- W3045872183 hasRelatedWork W6985466 @default.
- W3045872183 isParatext "false" @default.
- W3045872183 isRetracted "false" @default.
- W3045872183 magId "3045872183" @default.
- W3045872183 workType "article" @default.