Analysis of the BUT Diarization System for VoxConverse Challenge

TL;DR

This paper presents the BUT team's diarization system for VoxConverse, combining multiple processing steps and models, achieving top performance in the challenge metrics.

Contribution

The paper introduces a comprehensive diarization system with novel integration of clustering, Bayesian HMM, and overlapped speech handling for VoxConverse.

Findings

Achieved second place in diarization error rate

First place in Jaccard error rate

Detailed comparison of system components

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).