speech-1

Target Speaker Extraction

3260 papers • 126 benchmarks • 313 datasets

Extract the dialogue content of the specified target in a multi-person dialogue.

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in target-speaker-extraction-1

Trend

Dataset

Best Model

Actions

No benchmarks available.

Libraries

i

Use these libraries to find target-speaker-extraction-1 models and implementations

Datasets

No datasets available.

Subtasks

No subtasks available.

Most implemented papers

Target Speaker Verification With Selective Auditory Attention for Single and Multi-Talker Speech

Haizhou Li, Chenglin Xu, Wei Rao, Jibin Wu•Mon Mar 29 2021

A unified speaker verification framework for both single and multi-talker speech, that is able to pay selective auditory attention to the target speaker, that jointly optimizes a speaker attention module and a speaker representation module via multitask learning is proposed.

33

Content

0

Paper Graph

Selective Listening by Synchronizing Speech With Lips

Haizhou Li, Chenglin Xu, Ruijie Tao, Zexu Pan•Sun Jun 13 2021

This work proposes a self-supervised pre-training strategy, to exploit the speech-lip synchronization cue for target speaker extraction, which allows us to leverage abundant unlabeled in-domain data.

53 0

Paper Graph

L-SpEx: Localized Target Speaker Extraction

E. Chng, Chenglin Xu, Haizhou Li, Meng Ge, Longbiao Wang, J. Dang•Sun Feb 20 2022

This paper designs a speaker localizer driven by the target speaker’s embedding to extract the spatial features, including direction-of-arrival (DOA) of thetarget speaker and beamforming output, and proposes an end-to-end localized target speaker extraction on pure speech cues, that is called L-SpEx.

29 0

Paper Graph

A Hybrid Continuity Loss to Reduce Over-Suppression for Time-domain Target Speaker Extraction

Zexu Pan, Haizhou Li, Meng Ge•Wed Mar 30 2022

Experimental results show that the proposed loss function reduces the over-suppression and improves the word error rate of speech recognition on both clean and noisy two-speakers mixtures, without harming the re-constructed speech quality.

25 0

Paper Graph

ImagineNet: Target Speaker Extraction with Intermittent Visual Cue Through Embedding Inpainting

Zexu Pan, Haizhou Li, Wupeng Wang, Marvin Borsdorf•Sun Oct 30 2022

A joint speaker extraction and visual embedding inpainting framework to explore the mutual benefits and shows that the experimental results show that the proposed method outperforms the baseline in terms of signal quality, perceptual quality, and intelligibility.

15 0

Paper Graph

GPU-accelerated Guided Source Separation for Meeting Transcription

Daniel Povey, S. Khudanpur, Desh Raj•Fri Dec 09 2022

An improved implementation of GSS is described that leverages the power of modern GPU-based pipelines, including batched processing of frequencies and segments, to provide 300x speed-up over CPU-based inference.

47 0

Paper Graph

Adding a benchmark result helps the community track progress.