L-SpEx: Localized Target Speaker Extraction

TL;DR

L-SpEx introduces an end-to-end speech extraction method that localizes and extracts a target speaker using only speech cues, leveraging spatial features and attention mechanisms, and significantly improves performance in reverberant multi-channel environments.

Contribution

The paper presents a novel end-to-end localized speaker extraction model that does not rely on visual cues, utilizing a speaker localizer and spatial attention for improved accuracy.

Findings

Outperforms baseline systems on MC-Libri2Mix dataset.

Effectively estimates direction-of-arrival (DOA) and beamforming output.

Enhances target speaker extraction in reverberant multi-channel settings.

Abstract

Speaker extraction aims to extract the target speaker's voice from a multi-talker speech mixture given an auxiliary reference utterance. Recent studies show that speaker extraction benefits from the location or direction of the target speaker. However, these studies assume that the target speaker's location is known in advance or detected by an extra visual cue, e.g., face image or video. In this paper, we propose an end-to-end localized target speaker extraction on pure speech cues, that is called L-SpEx. Specifically, we design a speaker localizer driven by the target speaker's embedding to extract the spatial features, including direction-of-arrival (DOA) of the target speaker and beamforming output. Then, the spatial cues and target speaker's embedding are both used to form a top-down auditory attention to the target speaker. Experiments on the multi-channel reverberant dataset called MC-Libri2Mix show that our L-SpEx approach significantly outperforms the baseline system.