ABSP System for The Third DIHARD Challenge

TL;DR

This paper presents a speaker diarization system for the DIHARD III challenge that uses acoustic domain identification and domain-dependent clustering to improve diarization accuracy.

Contribution

The novel contribution is the development of an ADI system based on speaker embeddings and the integration of domain-dependent thresholds and PCA parameters for improved diarization.

Findings

Achieved a 9.63% relative DER reduction in core conditions.

Achieved a 10.64% relative DER reduction in full conditions.

Demonstrated effectiveness of domain-dependent processing in speaker diarization.

Abstract

This report describes the speaker diarization system developed by the ABSP Laboratory team for the third DIHARD speech diarization challenge. Our primary contribution is to develop acoustic domain identification (ADI) system for speaker diarization. We investigate speaker embeddings based ADI system. We apply a domain-dependent threshold for agglomerative hierarchical clustering. Besides, we optimize the parameters for PCA-based dimensionality reduction in a domain-dependent way. Our method of integrating domain-based processing schemes in the baseline system of the challenge achieved a relative improvement of $9.63\%$ and $10.64\%$ in DER for core and full conditions, respectively, for Track 1 of the DIHARD III evaluation set.