Matches in SemOpenAlex for { <https://semopenalex.org/work/W4214698081> ?p ?o ?g. }
- W4214698081 endingPage "1664" @default.
- W4214698081 startingPage "1650" @default.
- W4214698081 abstract "A speaker extraction algorithm seeks to extract the speech of a target speaker from a multi-talker speech mixture when given a cue that represents the target speaker, such as a pre-enrolled speech utterance, or an accompanying video track. Visual cues are particularly useful when a pre-enrolled speech is not available. In this work, we don’t rely on the target speaker’s pre-enrolled speech, but rather use the target speaker’s face track as the speaker cue, that is referred to as the auxiliary reference, to form an attractor towards the target speaker. We advocate that the temporal synchronization between the speech and its accompanying lip movements is a direct and dominant audio-visual cue. Therefore, we propose a self-supervised pre-training strategy, to exploit the speech-lip synchronization cue for target speaker extraction, which allows us to leverage abundant unlabeled in-domain data. We transfer the knowledge from the pre-trained model to the attractor encoder of the speaker extraction network. We show that the proposed speaker extraction network outperforms various competitive baselines in terms of signal quality, perceptual quality, and intelligibility, achieving state-of-the-art performance." @default.
- W4214698081 created "2022-03-02" @default.
- W4214698081 creator A5026034735 @default.
- W4214698081 creator A5032690182 @default.
- W4214698081 creator A5060530570 @default.
- W4214698081 creator A5080053477 @default.
- W4214698081 date "2022-01-01" @default.
- W4214698081 modified "2023-09-25" @default.
- W4214698081 title "Selective Listening by Synchronizing Speech With Lips" @default.
- W4214698081 cites W1482149378 @default.
- W4214698081 cites W1552314771 @default.
- W4214698081 cites W1966823466 @default.
- W4214698081 cites W1991139021 @default.
- W4214698081 cites W2015143272 @default.
- W4214698081 cites W2029199293 @default.
- W4214698081 cites W2069681747 @default.
- W4214698081 cites W2081144555 @default.
- W4214698081 cites W2098717567 @default.
- W4214698081 cites W2103104224 @default.
- W4214698081 cites W2107938580 @default.
- W4214698081 cites W2141411743 @default.
- W4214698081 cites W2147455188 @default.
- W4214698081 cites W2150769028 @default.
- W4214698081 cites W2167709189 @default.
- W4214698081 cites W2221409856 @default.
- W4214698081 cites W2333091651 @default.
- W4214698081 cites W2400006663 @default.
- W4214698081 cites W2521686623 @default.
- W4214698081 cites W2558649592 @default.
- W4214698081 cites W2587150483 @default.
- W4214698081 cites W2594690981 @default.
- W4214698081 cites W2604379605 @default.
- W4214698081 cites W2726515241 @default.
- W4214698081 cites W2734774145 @default.
- W4214698081 cites W2741151796 @default.
- W4214698081 cites W2800022361 @default.
- W4214698081 cites W2808631503 @default.
- W4214698081 cites W2844030168 @default.
- W4214698081 cites W2888968865 @default.
- W4214698081 cites W2890952074 @default.
- W4214698081 cites W2891833136 @default.
- W4214698081 cites W2900292050 @default.
- W4214698081 cites W2924115626 @default.
- W4214698081 cites W2938646939 @default.
- W4214698081 cites W2950864153 @default.
- W4214698081 cites W2951130829 @default.
- W4214698081 cites W2952218014 @default.
- W4214698081 cites W2959214850 @default.
- W4214698081 cites W2962788625 @default.
- W4214698081 cites W2962960500 @default.
- W4214698081 cites W2963082324 @default.
- W4214698081 cites W2963103134 @default.
- W4214698081 cites W2963528589 @default.
- W4214698081 cites W2964058413 @default.
- W4214698081 cites W2964171275 @default.
- W4214698081 cites W2964283370 @default.
- W4214698081 cites W2972460025 @default.
- W4214698081 cites W2972513594 @default.
- W4214698081 cites W2972568703 @default.
- W4214698081 cites W2973054998 @default.
- W4214698081 cites W2973062255 @default.
- W4214698081 cites W3007376164 @default.
- W4214698081 cites W3008400075 @default.
- W4214698081 cites W3010457696 @default.
- W4214698081 cites W3015199127 @default.
- W4214698081 cites W3015445830 @default.
- W4214698081 cites W3015623828 @default.
- W4214698081 cites W3015636705 @default.
- W4214698081 cites W3024147341 @default.
- W4214698081 cites W3036438531 @default.
- W4214698081 cites W3043140951 @default.
- W4214698081 cites W3095082129 @default.
- W4214698081 cites W3095379519 @default.
- W4214698081 cites W3096214032 @default.
- W4214698081 cites W3097653961 @default.
- W4214698081 cites W3097741049 @default.
- W4214698081 cites W3103434036 @default.
- W4214698081 cites W3105024549 @default.
- W4214698081 cites W3105928222 @default.
- W4214698081 cites W3116298410 @default.
- W4214698081 cites W3136499730 @default.
- W4214698081 cites W3148386121 @default.
- W4214698081 cites W3160672713 @default.
- W4214698081 cites W3162475350 @default.
- W4214698081 cites W3162534564 @default.
- W4214698081 cites W3163287738 @default.
- W4214698081 cites W3163573274 @default.
- W4214698081 cites W3182657421 @default.
- W4214698081 cites W3185109982 @default.
- W4214698081 cites W3190394743 @default.
- W4214698081 cites W3198234746 @default.
- W4214698081 cites W3198575250 @default.
- W4214698081 cites W3204420730 @default.
- W4214698081 cites W3211358952 @default.
- W4214698081 cites W3213412873 @default.
- W4214698081 cites W4205654098 @default.
- W4214698081 cites W4225682608 @default.
- W4214698081 cites W4289665794 @default.