CHiME-5
Emerging4papers using it
2023first seen
CHiME-5 is a dataset used to evaluate speech recognition systems in the context of highly-overlapped dinner party speech recorded by multiple microphone arrays.
Papers using CHiME-5 (4)
- Concurrent Speaker Detection: A multi-microphone Transformer-Based
ApproachThe CHiME-7 UDASE task: Unsupervised domain adaptation for
conversational speech enhancementAcoustic modeling for Overlapping Speech Recognition: JHU Chime-5
Challenge SystemObjective and subjective evaluation of speech enhancement methods in the
UDASE task of the 7th CHiME challenge