Third DIHARD Challenge Evaluation Plan
This is an incremental challenge for the speaker diarization community to benchmark and improve system robustness.
The paper introduces the third DIHARD challenge, which aims to enhance speaker diarization systems' robustness to variations in equipment, noise, and conversational domains, by evaluating performance through two tracks: one with reference speech segmentation and another from raw audio.
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.