Listening to Each Speaker One by One with Recurrent Selective Hearing Networks

By Keisuke Kinoshita, Lukas Drude, Marc Delcroix, Tomohiro Nakatani. 2018

82 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Signal Processing

Citations in the corpus

Single Channel Voice Separation for Unknown Number of Speakers Under Reverberant and Noisy Settings (2020, background): “The authors of [17, 18, 19] suggest to separate one speaker at a time using a recursive solution.”
What’s all the Fuss about Free Universal Sound Separation Data? (2020, background): “This paper provides three main contributions: In terms of variable source separation, Kinoshita et al. [8] proposed another approach for handling variable numbers of sources, where the separation network is made recurrent over sources.”
Sequence to Multi-Sequence Learning via Conditional Chain Mapping for Mixture Signals (2020, methodology): “It is worth mentioning that the upper bound of speaker number was set as 5 in our trained model, so there is a strong tendency to predict silence at the 6-th step (similar with the observation from RSAN [18]), leading to higher accuracy for 5-speaker mixtures than the 4-speaker ones.”
Encoder-Decoder Based Attractor Calculation for End-to-End Neural Diarization (2021, methodology): “For speech separation, RSAN [9], [10] and one-and-rest permutation invariant training (OR-PIT) [42] can be used.”
Multi-talker ASR for an unknown number of sources: Joint training of source counting, separation and ASR (2020, methodology): “A stopping criterion can, however, be integrated into the separator by using thresholding or an additional output flag, as inspired by [16].”
End-to-End Speaker Diarization for an Unknown Number of Speakers with Encoder-Decoder Based Attractors (2020, methodology): “One series of methods involve applying the one-vs-rest approach iteratively [18, 19, 20, 21].”
Neural Speaker Diarization with Speaker-Wise Chain Rule (2020, methodology): “Our study is inspired by a similar speaker-wise decoding model proposed for speech separation task [35], and it was applied in speaker diarization as a down-stream task of speech separation [36].”
Recursive speech separation for unknown number of speakers (2019, background): “e can design a system that focuses on separating some of the clearest speakers, that is, a few speakers who are close to the microphone. Recently, similar recursive separation approaches are proposed [21, 22]. However, their works focus on the case where the number of speakers are same or less than ”
Continuous Speech Separation with Conformer (2020, background): “In [17, 18, 19], the additional speaker identification module is introduced to enhance the separation performance for both perceptual and recognition.”