Showing eess.ASShow all
3 papers · 1 filter
eess.AS2026
DualSpecSE: A Dual-Path Speech Enhancement Network Integrating Mel and Complex Spectrograms
Xingchen Li, Ziqian Wang, Zikai Liu +4
In this paper, we propose DualSpecSE, a speech enhancement framework that jointly models Mel-spectrogram and complex spectrogram in a dual-path architecture for improved ASR perfor…
eess.AS2026
EvoTSE: Evolving Enrollment for Target Speaker Extraction
Zikai Liu, Ziqian Wang, Xingchen Li +4
Target Speaker Extraction (TSE) aims to isolate a specific speaker's voice from a mixture, guided by a pre-recorded enrollment. While TSE bypasses the global permutation ambiguity…
eess.AS2025
SenSE: Semantic-Aware High-Fidelity Universal Speech Enhancement
Xingchen Li, Hanke Xie, Ziqian Wang +4
Generative Universal Speech Enhancement (USE) methods aim to leverage generative models to improve speech quality under various types of distortions. However, existing generative s…