FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3132830522> ?p ?o ?g. }

• W3132830522 endingPage "101360" @default.
• W3132830522 startingPage "101360" @default.
• W3132830522 abstract "Multi-source localization is an important and challenging technique for multi-talker conversation analysis. This paper proposes a novel supervised learning method using deep neural networks to estimate the direction of arrival (DOA) of all the speakers simultaneously from the audio mixture. At the heart of the proposal is a source splitting mechanism that creates source-specific intermediate representations inside the network. This allows our model to give source-specific posteriors as the output unlike the traditional multi-label classification approach. Existing deep learning methods perform a frame level prediction, whereas our approach performs an utterance level prediction by incorporating temporal selection and averaging inside the network to avoid post-processing. We also experiment with various loss functions and show that a variant of earth mover distance (EMD) is very effective in classifying DOA at a very high resolution by modeling inter-class relationships. In addition to using the prediction error as a metric for evaluating our localization model, we also establish its potency as a frontend with automatic speech recognition (ASR) as the downstream task. We convert the estimated DOAs into a feature suitable for ASR and pass it as an additional input feature to a strong multi-channel and multi-talker speech recognition baseline. This added input feature drastically improves the ASR performance and gives a word error rate (WER) of 6.3% on the evaluation data of our simulated noisy two speaker mixtures, while the baseline which does not use explicit localization input has a WER of 11.5%. We also perform ASR evaluation on real recordings with the overlapped set of the MC-WSJ-AV corpus in addition to simulated mixtures. • An end-to-end multi-source direction of arrival (DOA) estimation method is proposed. • The method tackles multi-source DOA estimation via single-source DOA estimation. • The DOA classification network is optimized to learn inter-class relationships for a fine-grained angle resolution. • DOA estimation is used as a frontend for multi-talker automatic speech recognition." @default.
• W3132830522 created "2021-03-01" @default.
• W3132830522 creator A5001291873 @default.
• W3132830522 creator A5021176244 @default.
• W3132830522 creator A5022496760 @default.
• W3132830522 creator A5034476404 @default.
• W3132830522 creator A5048730867 @default.
• W3132830522 date "2022-09-01" @default.
• W3132830522 modified "2023-10-13" @default.
• W3132830522 title "Deep learning based multi-source localization with source splitting and its effectiveness in multi-talker speech recognition" @default.
• W3132830522 cites W2069432516 @default.
• W3132830522 cites W2069681747 @default.
• W3132830522 cites W2110059230 @default.
• W3132830522 cites W2113638573 @default.
• W3132830522 cites W2117678320 @default.
• W3132830522 cites W2128131274 @default.
• W3132830522 cites W2131090450 @default.
• W3132830522 cites W2141411743 @default.
• W3132830522 cites W2143668817 @default.
• W3132830522 cites W2147166770 @default.
• W3132830522 cites W2160783189 @default.
• W3132830522 cites W2164502538 @default.
• W3132830522 cites W2168729028 @default.
• W3132830522 cites W2242685705 @default.
• W3132830522 cites W2398776621 @default.
• W3132830522 cites W2403149086 @default.
• W3132830522 cites W2460742184 @default.
• W3132830522 cites W2526425061 @default.
• W3132830522 cites W2551990143 @default.
• W3132830522 cites W2592109325 @default.
• W3132830522 cites W2611943505 @default.
• W3132830522 cites W2763188033 @default.
• W3132830522 cites W2772736377 @default.
• W3132830522 cites W2810934215 @default.
• W3132830522 cites W2884797218 @default.
• W3132830522 cites W2885219692 @default.
• W3132830522 cites W2886180730 @default.
• W3132830522 cites W2888793942 @default.
• W3132830522 cites W2889061314 @default.
• W3132830522 cites W2890964092 @default.
• W3132830522 cites W2891054259 @default.
• W3132830522 cites W2891833136 @default.
• W3132830522 cites W2892163332 @default.
• W3132830522 cites W2897977894 @default.
• W3132830522 cites W2900582271 @default.
• W3132830522 cites W2909607850 @default.
• W3132830522 cites W2948091552 @default.
• W3132830522 cites W2962780374 @default.
• W3132830522 cites W2964342924 @default.
• W3132830522 cites W2972693890 @default.
• W3132830522 cites W2972818416 @default.
• W3132830522 cites W2982471419 @default.
• W3132830522 cites W3008283340 @default.
• W3132830522 cites W3008762051 @default.
• W3132830522 cites W3015834770 @default.
• W3132830522 cites W3016131876 @default.
• W3132830522 cites W3016252974 @default.
• W3132830522 cites W3020336359 @default.
• W3132830522 cites W3162358523 @default.
• W3132830522 cites W3163193264 @default.
• W3132830522 cites W2972725301 @default.
• W3132830522 doi "https://doi.org/10.1016/j.csl.2022.101360" @default.
• W3132830522 hasPublicationYear "2022" @default.
• W3132830522 type Work @default.
• W3132830522 sameAs 3132830522 @default.
• W3132830522 citedByCount "28" @default.
• W3132830522 countsByYear W31328305222021 @default.
• W3132830522 countsByYear W31328305222022 @default.
• W3132830522 countsByYear W31328305222023 @default.
• W3132830522 crossrefType "journal-article" @default.
• W3132830522 hasAuthorship W3132830522A5001291873 @default.
• W3132830522 hasAuthorship W3132830522A5021176244 @default.
• W3132830522 hasAuthorship W3132830522A5022496760 @default.
• W3132830522 hasAuthorship W3132830522A5034476404 @default.
• W3132830522 hasAuthorship W3132830522A5048730867 @default.
• W3132830522 hasBestOaLocation W31328305222 @default.
• W3132830522 hasConcept C108583219 @default.
• W3132830522 hasConcept C154945302 @default.
• W3132830522 hasConcept C204321447 @default.
• W3132830522 hasConcept C28490314 @default.
• W3132830522 hasConcept C41008148 @default.
• W3132830522 hasConceptScore W3132830522C108583219 @default.
• W3132830522 hasConceptScore W3132830522C154945302 @default.
• W3132830522 hasConceptScore W3132830522C204321447 @default.
• W3132830522 hasConceptScore W3132830522C28490314 @default.
• W3132830522 hasConceptScore W3132830522C41008148 @default.
• W3132830522 hasLocation W31328305221 @default.
• W3132830522 hasLocation W31328305222 @default.
• W3132830522 hasOpenAccess W3132830522 @default.
• W3132830522 hasPrimaryLocation W31328305221 @default.
• W3132830522 hasRelatedWork W2126887587 @default.
• W3132830522 hasRelatedWork W2731899572 @default.
• W3132830522 hasRelatedWork W2939353110 @default.
• W3132830522 hasRelatedWork W2989698750 @default.
• W3132830522 hasRelatedWork W3009238340 @default.
• W3132830522 hasRelatedWork W3107474891 @default.
• W3132830522 hasRelatedWork W3215138031 @default.
• W3132830522 hasRelatedWork W4321369474 @default.

• next