DualStream Contextual Fusion Network: Efficient Target Speaker Extraction by Leveraging Mixture and Enrollment Interactions

TL;DR

This paper introduces DCF-Net, a novel time-frequency domain network that leverages mixture and enrollment interactions for more accurate target speaker extraction, outperforming state-of-the-art methods.

Contribution

The paper proposes a DualStream Fusion Block to better utilize contextual and interaction information between mixture and enrollment signals in speaker extraction.

Findings

Achieves 21.6 dB SI-SDRi on benchmark dataset.

Reduces target confusion to 0.4%.

Demonstrates robustness in noisy and reverberant environments.

Abstract

Target speaker extraction focuses on extracting a target speech signal from an environment with multiple speakers by leveraging an enrollment. Existing methods predominantly rely on speaker embeddings obtained from the enrollment, potentially disregarding the contextual information and the internal interactions between the mixture and enrollment. In this paper, we propose a novel DualStream Contextual Fusion Network (DCF-Net) in the time-frequency (T-F) domain. Specifically, DualStream Fusion Block (DSFB) is introduced to obtain contextual information and capture the interactions between contextualized enrollment and mixture representation across both spatial and channel dimensions, and then rich and consistent representations are utilized to guide the extraction network for better extraction. Experimental results demonstrate that DCF-Net outperforms state-of-the-art (SOTA) methods, achieving a scale-invariant signal-to-distortion ratio improvement (SI-SDRi) of 21.6 dB on the benchmark dataset, and exhibits its robustness and effectiveness in both noise and reverberation scenarios. In addition, the wrong extraction results of our model, called target confusion problem, reduce to 0.4%, which highlights the potential of DCF-Net for practical applications.