The DKU-MSXF Diarization System For The Voxceleb Speaker Recognition Challenge 2023

Abstract

This paper describes the DKU-MSXF submission to track 4 of the VoxCeleb Speaker Recognition Challenge 2023 (VoxSRC-23). Our system pipeline contains voice activity detection, clustering-based diarization, overlapped speech detection, and target-speaker voice activity detection, where each procedure has a fused output from 3 sub-models. Finally, we fuse different clustering-based and TSVAD-based diarization systems using DOVER-Lap and achieve the 4.30% diarization error rate (DER), which ranks first place on track 4 of the challenge leaderboard.

Stats

• citations4
• S2 citations—
• github stars0
• HF likes0
• heat score5.24
• arxiv keycheng2023the

Related papers

• The DKU-MSXF Speaker Verification System For The Voxceleb Speaker Recognition Challenge 2023 (2023)0.00
• Microsoft Speaker Diarization System For The Voxceleb Speaker Recognition Challenge 2020 (2020)11.93
• The BUCEA Speaker Diarization System For The Voxceleb Speaker Recognition Challenge 2022 (2022)0.00
• The HUAWEI Speaker Diarisation System For The Voxceleb Speaker Diarisation Challenge (2020)0.00
• Analysis Of The BUT Diarization System For Voxconverse Challenge (2020)8.82
• North America Bixby Speaker Diarization System For The Voxceleb Speaker Recognition Challenge 2021 (2021)0.00
• Gist-aiter Speaker Diarization System For Voxceleb Speaker Recognition Challenge (voxsrc) 2023 (2023)0.00
• The Dku-duke-lenovo System Description For The Third DIHARD Speech Diarization Challenge (2021)0.00