Geodesic Interpolation of Frame-Wise Speaker Embeddings for the Diarization of Meeting Scenarios.
Tobias Cord-LandwehrChristoph BöddekerCatalin ZorilaRama DoddipatlaReinhold Haeb-UmbachPublished in: ICASSP (2024)
Keyphrases
- speaker diarization
- speech recognition
- speech activity detection
- bayesian information criterion
- pairwise
- low dimensional
- manifold learning
- real world
- frame rate
- linear interpolation
- broadcast news
- image frames
- euclidean space
- learning scenarios
- vector space
- reference frame
- video frames
- speaker identification
- multi modal
- speaker verification
- interpolation methods
- pattern recognition
- feature extraction
- meeting room
- image sequences