SAGRNN: Self-Attentive Gated RNN for Binaural Speaker Separation with Interaural Cue Preservation.
Ke TanBuye XuAnurag KumarEliya NachmaniYossi AdiPublished in: CoRR (2020)
Keyphrases
- sound source
- audio visual
- recurrent neural networks
- focus of attention
- visual motion
- speech signal
- nearest neighbor
- multi modal
- speech recognition
- automatic speech recognition
- visual attention
- speaker recognition
- speaker identification
- speaker verification
- visual data
- neural network
- visual information
- video sequences
- multimedia
- eye movements
- blind source separation
- optical flow
- image sequences
- computer vision
- multiple cues
- data sets
- pre attentive