Third DIHARD Challenge Evaluation Plan
arXiv:2006.05815
Abstract
This paper introduces the third DIHARD challenge, the third in a series of speaker diarization challenges intended to improve the robustness of diarization systems to variation in recording equipment, noise conditions, and conversational domain. The challenge comprises two tracks evaluating diarization performance when starting from a reference speech segmentation (track 1) and diarization from raw audio scratch (track 2). We describe the task, metrics, datasets, and evaluation protocol.
Version 1.2 - Planned schedule updated - Updated numbers in tables from final versions of development/evaluation sets - Corrected typo
Cited by in corpus (5)
- VoxSRC 2020: The Second VoxCeleb Speaker Recognition Challenge
- Separation Guided Speaker Diarization in Realistic Mismatched Conditions
- Online Streaming End-to-End Neural Diarization Handling Overlapping Speech and Flexible Numbers of Speakers
- The DKU-Duke-Lenovo System Description for the Third DIHARD Speech Diarization Challenge
- ABSP System for The Third DIHARD Challenge