Analysis of the but Diarization System for Voxconverse Challenge
Federico Landini, Ondrej Glembek, Pavel Matejka, Johan Rohdin, Lukás Burget, Mireia Díez, Anna Silnova
Abstract
This paper describes the system developed by the BUT team for the fourth track of the VoxCeleb Speaker Recognition Challenge, focusing on diarization on the VoxConverse dataset. The system consists of signal pre-processing, voice activity detection, speaker embedding extraction, an initial agglomerative hierarchical clustering followed by diarization using a Bayesian hidden Markov model, a reclustering step based on per-speaker global embeddings and overlapped speech detection and handling. We provide comparisons for each of the steps and share the implementation of the most relevant modules of our system. Our system scored second in the challenge in terms of the primary metric (diarization error rate) and first according to the secondary metric (Jaccard error rate).
BibTeX
@inproceedings{icassp2021_analysisofthebut,
title = {Analysis of the but Diarization System for Voxconverse Challenge},
author = {Federico Landini and Ondrej Glembek and Pavel Matejka and Johan Rohdin and Lukás Burget and Mireia Díez and Anna Silnova},
booktitle = {ICASSP 2021},
year = {2021}
}