Once more Diarization: Improving meeting transcription systems through segment-level speaker reassignment

TL;DR

This paper introduces a segment-level speaker reassignment method that improves meeting transcription accuracy by reducing speaker confusion errors, especially in noisy or overlapping speech conditions.

Contribution

It presents a novel approach of revisiting speaker attribution after speech enhancement to significantly reduce diarization errors in meeting transcription systems.

Findings

Reduces speaker confusion errors by at least 40%.

Effective across different system configurations and datasets.

Enhances diarization accuracy in noisy and overlapping speech scenarios.

Abstract

Diarization is a crucial component in meeting transcription systems to ease the challenges of speech enhancement and attribute the transcriptions to the correct speaker. Particularly in the presence of overlapping or noisy speech, these systems have problems reliably assigning the correct speaker labels, leading to a significant amount of speaker confusion errors. We propose to add segment-level speaker reassignment to address this issue. By revisiting, after speech enhancement, the speaker attribution for each segment, speaker confusion errors from the initial diarization stage are significantly reduced. Through experiments across different system configurations and datasets, we further demonstrate the effectiveness and applicability in various domains. Our results show that segment-level speaker reassignment successfully rectifies at least 40% of speaker confusion word errors, highlighting its potential for enhancing diarization accuracy in meeting transcription systems.