URGENT-PK: Perceptually-Aligned Ranking Model Designed for Speech Enhancement Competition
Abstract
The Mean Opinion Score (MOS) is fundamental to speech quality assessment. However, its acquisition requires significant human annotation. Although deep neural network approaches, such as DNSMOS and UTMOS, have been developed to predict MOS to avoid this issue, they often suffer from insufficient training data. Recognizing that the comparison of speech enhancement (SE) systems prioritizes a reliable system comparison over absolute scores, we propose URGENT-PK, a novel ranking approach leveraging pairwise comparisons. URGENT-PK takes homologous enhanced speech pairs as input to predict relative quality rankings. This pairwise paradigm efficiently utilizes limited training data, as all pairwise permutations of multiple systems constitute a training instance. Experiments across multiple open test sets demonstrate URGENT-PK's superior system-level ranking performance over state-of-the-art baselines, despite its simple network architecture and limited training data.
Keywords
Cite
@article{arxiv.2506.23874,
title = {URGENT-PK: Perceptually-Aligned Ranking Model Designed for Speech Enhancement Competition},
author = {Jiahe Wang and Chenda Li and Wei Wang and Wangyou Zhang and Samuele Cornell and Marvin Sach and Robin Scheibler and Kohei Saijo and Yihui Fu and Zhaoheng Ni and Anurag Kumar and Tim Fingscheidt and Shinji Watanabe and Yanmin Qian},
journal= {arXiv preprint arXiv:2506.23874},
year = {2025}
}
Comments
Submitted to ASRU2025