.png){kind=icon}
返回
So-DAS: A Two-Step Soft-Direction-Aware Speech Separation Framework
.png){kind=icon}
PRE
AI
DOI：10.1109/LSP.2023.3248952.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Most existing direction-aware speech separation systems lead to performance degradation when the angle difference between speakers is small due to the low spatial discrimination. To address this issue, we propose a two-step soft-direction-aware speech separation (So-DAS) framework, which consists of a direction of arrival (DOA) estimation module and a speech separation module. First, the two modules are individually optimized, and directional features (DFs) derived from ground-truth DOAs are utilized as spatial information to facilitate the separation module. Next, the two modules are cascaded and optimized with only separation loss, and the DFs are generated using the estimator outputs. By this means, the consistency between the two modules is strengthened, and thus spatial cues that are more beneficial to the separation task can be exploited by the network itself. The experimental results show that compared to the baselines, DFs extracted by our proposed method provides clearer superiority, especially when the angle difference between speakers is small. In addition, our approach yields a state-of-the-art word error rate of 3.4% on the real-recorded utterance-wise LibriCSS dataset.
Keyword:
Direction-of-arrival estimation
Particle separators
Training data
Training
Signal to noise ratio
Reflection
Optimization
Multi-channel speech separation
directional feature
deep learning
期刊
IF:
9.6
论文数:
1.1W
被引数:
1.7W
机构
引用论文
A Comparison of the Effectiveness of Sodium Stibogluconate Monotherapy to Sodium Stibogluconate and Paromomycin Combination for the Treatment of Severe Post Kala Azar Dermal Leishmaniasis in South Sudan – A Retrospective Cohort Study.png)
PLOS ONE
OAAI
PLOS ONEIF0
OAAI分享收藏Shape analysis of the neostriatum in frontotemporal lobar degeneration, Alzheimer's disease, and controlsNeuroImage
NeuroImageIF0
PREAI
分享收藏ChemInform Abstract: Thermodynamic Modelling of the Mg—Ge—Si, Mg—Ge—Sn, Mg—Pb—Si and Mg—Pb—Sn Systems.ChemInform
ChemInformIF0
PREAI
分享收藏Feature Selection and Overlapping Clustering-Based Multilabel Classification Model基于特征选择与重叠聚类的多标签分类模型
OAAI
OAAI分享收藏
查看更多内容