CASIA OpenIR

Browse/Search Results:  1-4 of 4 Help

Selected(0)Clear Items/Page:    Sort:
Learning Driving Models From Parallel End-to-End Driving Data Set 期刊论文
PROCEEDINGS OF THE IEEE, 2020, 卷号: 108, 期号: 2, 页码: 262-273
Authors:  Chen, Long;  Wang, Qing;  Lu, Xiankai;  Cao, Dongpu;  Wang, Fei-Yue
Favorite  |  View/Download:22/0  |  Submit date:2020/03/30
Data models  Training  Adaptation models  Task analysis  Reinforcement learning  Decision making  Transforms  Data set  end-to-end driving  parallel driving  
End-to-End Post-Filter for Speech Separation With Deep Attention Fusion Features 期刊论文
IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2020, 卷号: 28, 页码: 1303-1314
Authors:  Fan, Cunhang;  Tao, Jianhua;  Liu, Bin;  Yi, Jiangyan;  Wen, Zhengqi;  Liu, Xuefei
Favorite  |  View/Download:5/0  |  Submit date:2020/06/22
Feature extraction  Training  Interference  Speech enhancement  Clustering algorithms  Spectrogram  Speech separation  end-to-end post-filter  deep attention fusion features  deep clustering  permutation invariant training  
Forward-Backward Decoding Sequence for Regularizing End-to-End TTS 期刊论文
IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2019, 卷号: 27, 期号: 12, 页码: 2067-2079
Authors:  Zheng, Yibin;  Tao, Jianhua;  Wen, Zhengqi;  Yi, Jiangyan
Favorite  |  View/Download:12/0  |  Submit date:2020/03/30
Decoding  Training  Speech processing  Linguistics  Acoustics  Speech recognition  Forward-backward  regularization  encoder-decoder with attention  end-to-end  joint-training  TTS  
Real-time segmentation of various insulators using generative adversarial networks 期刊论文
IET COMPUTER VISION, 2018, 卷号: 12, 期号: 5, 页码: 596-602
Authors:  Chang, Wenkai;  Yang, Guodong;  Yu, Junzhi;  Liang, Zize
Favorite  |  View/Download:29/0  |  Submit date:2019/12/16
image segmentation  insulators  neural nets  power engineering computing  real-time pixel-level segmentation  generative adversarial networks  insulator segmentation algorithm  cluttered background  artificial thresholds  compact end-to-end neural network  visual saliency map  proposed two-stage training  segmentation quality