Matches in SemOpenAlex for { <https://semopenalex.org/work/W3210134537> ?p ?o ?g. }
- W3210134537 abstract "While permutation invariant training (PIT) based continuous speech separation (CSS) significantly improves the conversation transcription accuracy, it often suffers from speech leakages and failures in separation at hot spot regions because it has a fixed number of output channels. In this paper, we propose to apply recurrent selective attention network (RSAN) to CSS, which generates a variable number of output channels based on active speaker counting. In addition, we propose a novel block-wise dependency extension of RSAN by introducing dependencies between adjacent processing blocks in the CSS framework. It enables the network to utilize the separation results from the previous blocks to facilitate the current block processing. Experimental results on the LibriCSS dataset show that the RSAN-based CSS (RSAN-CSS) network consistently improves the speech recognition accuracy over PIT-based models. The proposed block-wise dependency modeling further boosts the performance of RSAN-CSS." @default.
- W3210134537 created "2021-11-08" @default.
- W3210134537 creator A5017825677 @default.
- W3210134537 creator A5029670581 @default.
- W3210134537 creator A5032260110 @default.
- W3210134537 creator A5040405220 @default.
- W3210134537 creator A5040519711 @default.
- W3210134537 creator A5051392570 @default.
- W3210134537 creator A5062250845 @default.
- W3210134537 date "2022-05-23" @default.
- W3210134537 modified "2023-10-11" @default.
- W3210134537 title "Continuous Speech Separation with Recurrent Selective Attention Network" @default.
- W3210134537 cites W1494198834 @default.
- W3210134537 cites W2030486566 @default.
- W3210134537 cites W2117678320 @default.
- W3210134537 cites W2221409856 @default.
- W3210134537 cites W2460742184 @default.
- W3210134537 cites W2734774145 @default.
- W3210134537 cites W2803322398 @default.
- W3210134537 cites W2889698889 @default.
- W3210134537 cites W2891054259 @default.
- W3210134537 cites W2891644395 @default.
- W3210134537 cites W2952218014 @default.
- W3210134537 cites W2972460025 @default.
- W3210134537 cites W2972956184 @default.
- W3210134537 cites W2980451698 @default.
- W3210134537 cites W3000358149 @default.
- W3210134537 cites W3015199127 @default.
- W3210134537 cites W3016232124 @default.
- W3210134537 cites W3020336359 @default.
- W3210134537 cites W3032514799 @default.
- W3210134537 cites W3116025434 @default.
- W3210134537 cites W3163652268 @default.
- W3210134537 cites W3163842642 @default.
- W3210134537 cites W3171278394 @default.
- W3210134537 cites W3185109982 @default.
- W3210134537 doi "https://doi.org/10.1109/icassp43922.2022.9746394" @default.
- W3210134537 hasPublicationYear "2022" @default.
- W3210134537 type Work @default.
- W3210134537 sameAs 3210134537 @default.
- W3210134537 citedByCount "2" @default.
- W3210134537 countsByYear W32101345372022 @default.
- W3210134537 countsByYear W32101345372023 @default.
- W3210134537 crossrefType "proceedings-article" @default.
- W3210134537 hasAuthorship W3210134537A5017825677 @default.
- W3210134537 hasAuthorship W3210134537A5029670581 @default.
- W3210134537 hasAuthorship W3210134537A5032260110 @default.
- W3210134537 hasAuthorship W3210134537A5040405220 @default.
- W3210134537 hasAuthorship W3210134537A5040519711 @default.
- W3210134537 hasAuthorship W3210134537A5051392570 @default.
- W3210134537 hasAuthorship W3210134537A5062250845 @default.
- W3210134537 hasBestOaLocation W32101345372 @default.
- W3210134537 hasConcept C119857082 @default.
- W3210134537 hasConcept C121332964 @default.
- W3210134537 hasConcept C138885662 @default.
- W3210134537 hasConcept C154945302 @default.
- W3210134537 hasConcept C19768560 @default.
- W3210134537 hasConcept C21308566 @default.
- W3210134537 hasConcept C24890656 @default.
- W3210134537 hasConcept C2524010 @default.
- W3210134537 hasConcept C2776061190 @default.
- W3210134537 hasConcept C2777200299 @default.
- W3210134537 hasConcept C2777210771 @default.
- W3210134537 hasConcept C28490314 @default.
- W3210134537 hasConcept C33923547 @default.
- W3210134537 hasConcept C41008148 @default.
- W3210134537 hasConcept C41895202 @default.
- W3210134537 hasConcept C61328038 @default.
- W3210134537 hasConceptScore W3210134537C119857082 @default.
- W3210134537 hasConceptScore W3210134537C121332964 @default.
- W3210134537 hasConceptScore W3210134537C138885662 @default.
- W3210134537 hasConceptScore W3210134537C154945302 @default.
- W3210134537 hasConceptScore W3210134537C19768560 @default.
- W3210134537 hasConceptScore W3210134537C21308566 @default.
- W3210134537 hasConceptScore W3210134537C24890656 @default.
- W3210134537 hasConceptScore W3210134537C2524010 @default.
- W3210134537 hasConceptScore W3210134537C2776061190 @default.
- W3210134537 hasConceptScore W3210134537C2777200299 @default.
- W3210134537 hasConceptScore W3210134537C2777210771 @default.
- W3210134537 hasConceptScore W3210134537C28490314 @default.
- W3210134537 hasConceptScore W3210134537C33923547 @default.
- W3210134537 hasConceptScore W3210134537C41008148 @default.
- W3210134537 hasConceptScore W3210134537C41895202 @default.
- W3210134537 hasConceptScore W3210134537C61328038 @default.
- W3210134537 hasLocation W32101345371 @default.
- W3210134537 hasLocation W32101345372 @default.
- W3210134537 hasOpenAccess W3210134537 @default.
- W3210134537 hasPrimaryLocation W32101345371 @default.
- W3210134537 hasRelatedWork W1511346092 @default.
- W3210134537 hasRelatedWork W1527532029 @default.
- W3210134537 hasRelatedWork W1968552888 @default.
- W3210134537 hasRelatedWork W2281307425 @default.
- W3210134537 hasRelatedWork W2374116601 @default.
- W3210134537 hasRelatedWork W2378167147 @default.
- W3210134537 hasRelatedWork W2464405057 @default.
- W3210134537 hasRelatedWork W2772323916 @default.
- W3210134537 hasRelatedWork W3210777354 @default.
- W3210134537 hasRelatedWork W3093134843 @default.
- W3210134537 isParatext "false" @default.
- W3210134537 isRetracted "false" @default.