Matches in SemOpenAlex for { <https://semopenalex.org/work/W3162341667> ?p ?o ?g. }
- W3162341667 abstract "Speech separation algorithms are often used to separate the target speech from other interfering sources. However, purely neural network based speech separation systems often cause nonlinear distortion that is harmful for automatic speech recognition (ASR) systems. The conventional mask-based minimum variance distortionless response (MVDR) beamformer can be used to minimize the distortion, but comes with high level of residual noise. Furthermore, the matrix operations (e.g., matrix inversion) involved in the conventional MVDR solution are sometimes numerically unstable when jointly trained with neural networks. In this paper, we propose a novel all deep learning MVDR framework, where the matrix inversion and eigenvalue decomposition are replaced by two recurrent neural networks (RNNs), to resolve both issues at the same time. The proposed method can greatly reduce the residual noise while keeping the target speech undistorted by leveraging on the RNN-predicted frame-wise beamforming weights. The system is evaluated on a Mandarin audio-visual corpus and compared against several state-of-the-art (SOTA) speech separation systems. Experimental results demonstrate the superiority of the proposed method across several objective metrics and ASR accuracy." @default.
- W3162341667 created "2021-05-24" @default.
- W3162341667 creator A5010277066 @default.
- W3162341667 creator A5022496760 @default.
- W3162341667 creator A5034476404 @default.
- W3162341667 creator A5056567731 @default.
- W3162341667 creator A5072495935 @default.
- W3162341667 creator A5078783442 @default.
- W3162341667 date "2021-06-06" @default.
- W3162341667 modified "2023-10-18" @default.
- W3162341667 title "ADL-MVDR: All Deep Learning MVDR Beamformer for Target Speech Separation" @default.
- W3162341667 cites W1552314771 @default.
- W3162341667 cites W1997270899 @default.
- W3162341667 cites W2058079016 @default.
- W3162341667 cites W2060108923 @default.
- W3162341667 cites W2120608389 @default.
- W3162341667 cites W2127851351 @default.
- W3162341667 cites W2130317597 @default.
- W3162341667 cites W2398042854 @default.
- W3162341667 cites W2402526332 @default.
- W3162341667 cites W2517616541 @default.
- W3162341667 cites W2714487941 @default.
- W3162341667 cites W2803583024 @default.
- W3162341667 cites W2890553422 @default.
- W3162341667 cites W2891378882 @default.
- W3162341667 cites W2909607850 @default.
- W3162341667 cites W2939497224 @default.
- W3162341667 cites W2952218014 @default.
- W3162341667 cites W2962935966 @default.
- W3162341667 cites W3011424113 @default.
- W3162341667 cites W3096008106 @default.
- W3162341667 cites W3097173744 @default.
- W3162341667 cites W3097213357 @default.
- W3162341667 cites W3097966334 @default.
- W3162341667 cites W3101330598 @default.
- W3162341667 doi "https://doi.org/10.1109/icassp39728.2021.9413594" @default.
- W3162341667 hasPublicationYear "2021" @default.
- W3162341667 type Work @default.
- W3162341667 sameAs 3162341667 @default.
- W3162341667 citedByCount "39" @default.
- W3162341667 countsByYear W31623416672020 @default.
- W3162341667 countsByYear W31623416672021 @default.
- W3162341667 countsByYear W31623416672022 @default.
- W3162341667 countsByYear W31623416672023 @default.
- W3162341667 crossrefType "proceedings-article" @default.
- W3162341667 hasAuthorship W3162341667A5010277066 @default.
- W3162341667 hasAuthorship W3162341667A5022496760 @default.
- W3162341667 hasAuthorship W3162341667A5034476404 @default.
- W3162341667 hasAuthorship W3162341667A5056567731 @default.
- W3162341667 hasAuthorship W3162341667A5072495935 @default.
- W3162341667 hasAuthorship W3162341667A5078783442 @default.
- W3162341667 hasBestOaLocation W31623416672 @default.
- W3162341667 hasConcept C11413529 @default.
- W3162341667 hasConcept C121332964 @default.
- W3162341667 hasConcept C126780896 @default.
- W3162341667 hasConcept C147168706 @default.
- W3162341667 hasConcept C153180895 @default.
- W3162341667 hasConcept C154945302 @default.
- W3162341667 hasConcept C155512373 @default.
- W3162341667 hasConcept C158693339 @default.
- W3162341667 hasConcept C163294075 @default.
- W3162341667 hasConcept C169756996 @default.
- W3162341667 hasConcept C194257627 @default.
- W3162341667 hasConcept C2776182073 @default.
- W3162341667 hasConcept C2776257435 @default.
- W3162341667 hasConcept C2776864781 @default.
- W3162341667 hasConcept C28490314 @default.
- W3162341667 hasConcept C41008148 @default.
- W3162341667 hasConcept C50644808 @default.
- W3162341667 hasConcept C54197355 @default.
- W3162341667 hasConcept C62520636 @default.
- W3162341667 hasConcept C76155785 @default.
- W3162341667 hasConceptScore W3162341667C11413529 @default.
- W3162341667 hasConceptScore W3162341667C121332964 @default.
- W3162341667 hasConceptScore W3162341667C126780896 @default.
- W3162341667 hasConceptScore W3162341667C147168706 @default.
- W3162341667 hasConceptScore W3162341667C153180895 @default.
- W3162341667 hasConceptScore W3162341667C154945302 @default.
- W3162341667 hasConceptScore W3162341667C155512373 @default.
- W3162341667 hasConceptScore W3162341667C158693339 @default.
- W3162341667 hasConceptScore W3162341667C163294075 @default.
- W3162341667 hasConceptScore W3162341667C169756996 @default.
- W3162341667 hasConceptScore W3162341667C194257627 @default.
- W3162341667 hasConceptScore W3162341667C2776182073 @default.
- W3162341667 hasConceptScore W3162341667C2776257435 @default.
- W3162341667 hasConceptScore W3162341667C2776864781 @default.
- W3162341667 hasConceptScore W3162341667C28490314 @default.
- W3162341667 hasConceptScore W3162341667C41008148 @default.
- W3162341667 hasConceptScore W3162341667C50644808 @default.
- W3162341667 hasConceptScore W3162341667C54197355 @default.
- W3162341667 hasConceptScore W3162341667C62520636 @default.
- W3162341667 hasConceptScore W3162341667C76155785 @default.
- W3162341667 hasLocation W31623416671 @default.
- W3162341667 hasLocation W31623416672 @default.
- W3162341667 hasOpenAccess W3162341667 @default.
- W3162341667 hasPrimaryLocation W31623416671 @default.
- W3162341667 hasRelatedWork W2118022799 @default.
- W3162341667 hasRelatedWork W2144114631 @default.
- W3162341667 hasRelatedWork W2251439150 @default.
- W3162341667 hasRelatedWork W2636483419 @default.