Matches in SemOpenAlex for { <https://semopenalex.org/work/W2787944098> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W2787944098 abstract "Several end-to-end deep learning approaches have been recently presented which extract either audio or visual features from the input images or audio signals and perform speech recognition. However, research on end-to-end audiovisual models is very limited. In this work, we present an end-to-end audiovisual model based on residual networks and Bidirectional Gated Recurrent Units (BGRUs). To the best of our knowledge, this is the first audiovisual fusion model which simultaneously learns to extract features directly from the image pixels and audio waveforms and performs within-context word recognition on a large publicly available dataset (LRW). The model consists of two streams, one for each modality, which extract features directly from mouth regions and raw waveforms. The temporal dynamics in each stream/modality are modeled by a 2-layer BGRU and the fusion of multiple streams/modalities takes place via another 2-layer BGRU. A slight improvement in the classification rate over an end-to-end audio-only and MFCC-based model is reported in clean audio conditions and low levels of noise. In presence of high levels of noise, the end-to-end audiovisual model significantly outperforms both audio-only models." @default.
- W2787944098 created "2018-03-06" @default.
- W2787944098 creator A5001790767 @default.
- W2787944098 creator A5009475700 @default.
- W2787944098 creator A5016033078 @default.
- W2787944098 creator A5024224610 @default.
- W2787944098 creator A5031069397 @default.
- W2787944098 creator A5061939508 @default.
- W2787944098 date "2018-02-18" @default.
- W2787944098 modified "2023-09-27" @default.
- W2787944098 title "End-to-end Audiovisual Speech Recognition" @default.
- W2787944098 hasPublicationYear "2018" @default.
- W2787944098 type Work @default.
- W2787944098 sameAs 2787944098 @default.
- W2787944098 citedByCount "3" @default.
- W2787944098 countsByYear W27879440982018 @default.
- W2787944098 countsByYear W27879440982019 @default.
- W2787944098 crossrefType "posted-content" @default.
- W2787944098 hasAuthorship W2787944098A5001790767 @default.
- W2787944098 hasAuthorship W2787944098A5009475700 @default.
- W2787944098 hasAuthorship W2787944098A5016033078 @default.
- W2787944098 hasAuthorship W2787944098A5024224610 @default.
- W2787944098 hasAuthorship W2787944098A5031069397 @default.
- W2787944098 hasAuthorship W2787944098A5061939508 @default.
- W2787944098 hasConcept C115961682 @default.
- W2787944098 hasConcept C151730666 @default.
- W2787944098 hasConcept C153180895 @default.
- W2787944098 hasConcept C154945302 @default.
- W2787944098 hasConcept C157968479 @default.
- W2787944098 hasConcept C204201278 @default.
- W2787944098 hasConcept C2779343474 @default.
- W2787944098 hasConcept C2780226545 @default.
- W2787944098 hasConcept C28490314 @default.
- W2787944098 hasConcept C41008148 @default.
- W2787944098 hasConcept C61328038 @default.
- W2787944098 hasConcept C74296488 @default.
- W2787944098 hasConcept C86803240 @default.
- W2787944098 hasConcept C99498987 @default.
- W2787944098 hasConceptScore W2787944098C115961682 @default.
- W2787944098 hasConceptScore W2787944098C151730666 @default.
- W2787944098 hasConceptScore W2787944098C153180895 @default.
- W2787944098 hasConceptScore W2787944098C154945302 @default.
- W2787944098 hasConceptScore W2787944098C157968479 @default.
- W2787944098 hasConceptScore W2787944098C204201278 @default.
- W2787944098 hasConceptScore W2787944098C2779343474 @default.
- W2787944098 hasConceptScore W2787944098C2780226545 @default.
- W2787944098 hasConceptScore W2787944098C28490314 @default.
- W2787944098 hasConceptScore W2787944098C41008148 @default.
- W2787944098 hasConceptScore W2787944098C61328038 @default.
- W2787944098 hasConceptScore W2787944098C74296488 @default.
- W2787944098 hasConceptScore W2787944098C86803240 @default.
- W2787944098 hasConceptScore W2787944098C99498987 @default.
- W2787944098 hasLocation W27879440981 @default.
- W2787944098 hasOpenAccess W2787944098 @default.
- W2787944098 hasPrimaryLocation W27879440981 @default.
- W2787944098 hasRelatedWork W1869491686 @default.
- W2787944098 hasRelatedWork W2039679570 @default.
- W2787944098 hasRelatedWork W2063425879 @default.
- W2787944098 hasRelatedWork W2076462394 @default.
- W2787944098 hasRelatedWork W2182448868 @default.
- W2787944098 hasRelatedWork W2594690981 @default.
- W2787944098 hasRelatedWork W2754826374 @default.
- W2787944098 hasRelatedWork W2788241093 @default.
- W2787944098 hasRelatedWork W2793122029 @default.
- W2787944098 hasRelatedWork W2897067191 @default.
- W2787944098 hasRelatedWork W2952746495 @default.
- W2787944098 hasRelatedWork W2959214850 @default.
- W2787944098 hasRelatedWork W2963192365 @default.
- W2787944098 hasRelatedWork W2980600748 @default.
- W2787944098 hasRelatedWork W3009620048 @default.
- W2787944098 hasRelatedWork W3121029767 @default.
- W2787944098 hasRelatedWork W3130492967 @default.
- W2787944098 hasRelatedWork W1833616679 @default.
- W2787944098 hasRelatedWork W2572763611 @default.
- W2787944098 hasRelatedWork W3129434801 @default.
- W2787944098 isParatext "false" @default.
- W2787944098 isRetracted "false" @default.
- W2787944098 magId "2787944098" @default.
- W2787944098 workType "article" @default.