Speaker segmentation using i-vector in meetings domain.
Leonardo Valeriano NeriHector N. B. PinheiroTsang Ing RenGeorge D. C. CavalcantiAndré Gustavo AdamiPublished in: ICASSP (2017)
Keyphrases
- audio visual
- segmentation algorithm
- domain specific
- segmentation method
- speaker diarization
- image segmentation
- object segmentation
- level set
- segmentation accuracy
- neural network
- medical images
- speech recognition
- domain independent
- shape prior
- multiscale
- region growing
- optimal segmentation
- brain mri
- speaker verification
- video sequences
- vector space
- ground truth
- multi modal
- maximum likelihood