In contrast to conventional speaker diarization methods that partition audio streams into segments, this project aims to develop robust techniques for segmenting transcripts into utterances by accurately distinguishing the speakers involved.
Director
- Jinho Choi - Associate Professor at Emory University
Publication
- Aligning Speakers: Evaluating and Visualizing Text-based Speaker Diarization Using Efficient Multiple Sequence Alignment. Gong, C.; Wu, P.; and Choi, J. D. Proceedings of the IEEE International Conference on Tools with Artificial Intelligence (ICTAI), 2023.
- Discriminative Speech Recognition Rescoring with Bert. Xu, L.; Gu, Y.; Kolehmainen, J.; Khan, H.; Gandhe, A.; Rastrow, A.; Stolcke, A.; Bulyko, I. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022.