Ensemble environment modeling using affine transform group

Tsao, Yu; Lin, Payton; Hu, Ting-Yao; Lu, Xugang

doi:10.1016/j.specom.2014.12.007

Cited by 3 publications

(1 citation statement)

References 67 publications

(72 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the online phase, an input reverberant speech is first processed by all dereverberation models simultaneously, and the outputs are integrated to ultimately generate the anechoic signals. The ensemble learning strategy, which has been proven to be able to improve system performance in speech enhancement [25] and ASR [26,27], is adopted in the task to increase the generalization ability of DDAEs. As will be introduced in the re- sults of experiments, conducted using the Mandarin hearing in noise test (MHINT) [28], a DDAE-based dereverberation system achieves the best quality and intelligibility scores when the training and testing conditions are similar (matched condition).…”

Section: Introductionmentioning

confidence: 99%

Speech Dereverberation Based on Integrated Deep and Ensemble Learning Algorithm

Lee

Wang

Chen

et al. 2018

2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

Reverberation, which is generally caused by sound reflections from walls, ceilings, and floors, can result in severe performance degradation of acoustic applications. Due to a complicated combination of attenuation and time-delay effects, the reverberation property is difficult to characterize, and it remains a challenging task to effectively retrieve the anechoic speech signals from reverberation ones. In the present study, we proposed a novel integrated deep and ensemble learning algorithm (IDEA) for speech dereverberation. The IDEA consists of offline and online phases. In the offline phase, we train multiple dereverberation models, each aiming to precisely dereverb speech signals in a particular acoustic environment; then a unified fusion function is estimated that aims to integrate the information of multiple dereverberation models. In the online phase, an input utterance is first processed by each of the dereverberation models. The outputs of all models are integrated accordingly to generate the final anechoic signal. We evaluated the IDEA on designed acoustic environments, including both matched and mismatched conditions of the training and testing data. Experimental results confirm that the proposed IDEA outperforms single deep-neural-network-based dereverberation model with the same model architecture and training data.

show abstract