Y-Vector: Multiscale Waveform Encoder for Speaker Embedding.
Ge ZhuFei JiangZhiyao DuanPublished in: Interspeech (2021)
Keyphrases
- multiscale
- vector space
- scale space
- image processing
- rate distortion
- speech recognition
- speaker verification
- image representation
- audio visual
- bit rate
- low complexity
- affine invariant
- speaker recognition
- speaker identification
- multiscale analysis
- coarse to fine
- video compression
- speaker diarization
- information hiding
- video sequences
- data hiding
- image formation
- wavelet coefficients
- frequency domain
- edge detection
- motion estimation
- feature vectors