The USTC-NERCSLIP Systems for the CHiME-7 DASR Challenge

TL;DR

This paper presents a system for speaker diarization and speech recognition in complex multi-speaker scenarios, achieving significant WER improvements through end-to-end models and spatial information rectification.

Contribution

The authors developed an end-to-end diarization system with a novel spatial rectification strategy and leveraged pre-trained models for speech recognition, improving performance on CHiME-7.

Findings

Achieved a DA-WER of 21.01% on CHiME-7 evaluation set

Reduced WER by 62.04% relative to baseline

Implemented multi-channel spatial information rectification

Abstract

This technical report details our submission system to the CHiME-7 DASR Challenge, which focuses on speaker diarization and speech recognition under complex multi-speaker scenarios. Additionally, it also evaluates the efficiency of systems in handling diverse array devices. To address these issues, we implemented an end-to-end speaker diarization system and introduced a rectification strategy based on multi-channel spatial information. This approach significantly diminished the word error rates (WER). In terms of recognition, we utilized publicly available pre-trained models as the foundational models to train our end-to-end speech recognition models. Our system attained a Macro-averaged diarization-attributed WER (DA-WER) of 21.01% on the CHiME-7 evaluation set, which signifies a relative improvement of 62.04% over the official baseline system.