X
Xugang Lu
Researcher at National Institute of Information and Communications Technology
Publications - 160
Citations - 3353
Xugang Lu is an academic researcher from National Institute of Information and Communications Technology. The author has contributed to research in topics: Speech enhancement & Intelligibility (communication). The author has an hindex of 20, co-authored 149 publications receiving 2588 citations. Previous affiliations of Xugang Lu include Japan Advanced Institute of Science and Technology.
Papers
More filters
Proceedings ArticleDOI
Speech enhancement based on deep denoising autoencoder.
TL;DR: Experimental results show that adding depth of the DAE consistently increase the performance when a large training data set is given, and compared with a minimum mean square error based speech enhancement algorithm, the proposed denoising DAE provided superior performance on the three objective evaluations.
Journal ArticleDOI
End-to-End Waveform Utterance Enhancement for Direct Evaluation Metrics Optimization by Fully Convolutional Neural Networks
TL;DR: In this paper, an end-to-end utterance-based speech enhancement framework using fully convolutional neural networks (FCN) was proposed to reduce the gap between the model optimization and the evaluation criterion.
Proceedings ArticleDOI
SNR-Aware Convolutional Neural Network Modeling for Speech Enhancement.
Szu-Wei Fu,Yu Tsao,Xugang Lu +2 more
TL;DR: CNN with the two proposed SNR-aware algorithms outperform the deep neural network counterpart in terms of standardized objective evaluations when using the same number of layers and nodes, suggesting their promising generalization capability for real-world applications.
Proceedings ArticleDOI
Complex spectrogram enhancement by convolutional neural network with multi-metrics learning
TL;DR: In this paper, a CNN model was proposed to estimate clean real and imaginary (RI) spectrograms from noisy ones, which are then used to synthesize enhanced speech waveforms.
Proceedings ArticleDOI
Raw waveform-based speech enhancement by fully convolutional networks
TL;DR: The proposed fully convolutional network (FCN) model can not only effectively recover the waveforms but also outperform the LPS- based DNN baseline in terms of short-time objective intelligibility (STOI) and perceptual evaluation of speech quality (PESQ).