Separating Long-Form Speech with Group-wise Permutation Invariant Training

Zhang, Wangyou; Chen, Zhuo; Kanda, Naoyuki; Liu, Shujie; Li, Jinyu; Eskimez, Şefik Emre; Yoshioka, Takuya; Xiao, Xiong; Meng, Ziyang; Ye, Qian; Wei, Furu

doi:10.21437/interspeech.2022-10362

Cited by 4 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many studies have been proposed to improve different aspects of the CSS framework [3,4,5,6,7,8,9,10]. We introduced a modulation factor based on segment overlap ratio to dynamically adjust the separation loss [3].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Location-Based Training for Multi-Channel Talker-Independent Speaker Separation

Taherian¹,

Tan²,

Wang³

2022

ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

The performance of automatic speech recognition (ASR) systems severely degrades when multi-talker speech overlap occurs. In meeting environments, speech separation is typically performed to improve the robustness of ASR systems. Recently, location-based training (LBT) was proposed as a new training criterion for multi-channel talker-independent speaker separation. Assuming fixed array geometry, LBT outperforms widely-used permutation-invariant training in fully overlapped utterances and matched reverberant conditions. This paper extends LBT to conversational multi-channel speaker separation. We introduce multi-resolution LBT to estimate the complex spectrograms from low to high time and frequency resolutions. With multi-resolution LBT, convolutional kernels are assigned consistently based on speaker locations in physical space. Evaluation results show that multi-resolution LBT consistently outperforms other competitive methods on the recorded LibriCSS corpus.

show abstract

Section: Introductionmentioning

confidence: 99%

“…In [4], a recurrent selective attention network is used to separate one speaker at a time. The work in [5] and [6] proposed new training criteria that generalizes PIT to capture long speech contexts. Li et al [7] proposed a dual-path separation model to leverage inter-segment information from a memory embedding pool.…”

Section: Introductionmentioning

confidence: 99%