Search

Your search keyword '"Li, Jingdong"' showing total 863 results
Start Over Author "Li, Jingdong"
863 results on '"Li, Jingdong"'

10. The PCG-AIID System for L3DAS22 Challenge: MIMO and MISO convolutional recurrent Network for Multi Channel Speech Enhancement and Speech Recognition

Author

Li, Jingdong, Zhu, Yuanyuan, Luo, Dawei, Liu, Yun, Cui, Guohui, and Li, Zhaoxia

Subjects
Electrical Engineering and Systems Science - Audio and Speech Processing, Computer Science - Sound
Abstract

This paper described the PCG-AIID system for L3DAS22 challenge in Task 1: 3D speech enhancement in office reverberant environment. We proposed a two-stage framework to address multi-channel speech denoising and dereverberation. In the first stage, a multiple input and multiple output (MIMO) network is applied to remove background noise while maintaining the spatial characteristics of multi-channel signals. In the second stage, a multiple input and single output (MISO) network is applied to enhance the speech from desired direction and post-filtering. As a result, our system ranked 3rd place in ICASSP2022 L3DAS22 challenge and significantly outperforms the baseline system, while achieving 3.2% WER and 0.972 STOI on the blind test-set., Comment: To appear at ICASSP 2022 (Accepted)

Published
2022

12. Uformer: A Unet based dilated complex & real dual-path conformer network for simultaneous speech enhancement and dereverberation

Author

Fu, Yihui, Liu, Yun, Li, Jingdong, Luo, Dawei, Lv, Shubo, Jv, Yukai, and Xie, Lei

Subjects
Electrical Engineering and Systems Science - Audio and Speech Processing, Computer Science - Sound
Abstract

Complex spectrum and magnitude are considered as two major features of speech enhancement and dereverberation. Traditional approaches always treat these two features separately, ignoring their underlying relationship. In this paper, we propose Uformer, a Unet based dilated complex & real dual-path conformer network in both complex and magnitude domain for simultaneous speech enhancement and dereverberation. We exploit time attention (TA) and dilated convolution (DC) to leverage local and global contextual information and frequency attention (FA) to model dimensional information. These three sub-modules contained in the proposed dilated complex & real dual-path conformer module effectively improve the speech enhancement and dereverberation performance. Furthermore, hybrid encoder and decoder are adopted to simultaneously model the complex spectrum and magnitude and promote the information interaction between two domains. Encoder decoder attention is also applied to enhance the interaction between encoder and decoder. Our experimental results outperform all SOTA time and complex domain models objectively and subjectively. Specifically, Uformer reaches 3.6032 DNSMOS on the blind test set of Interspeech 2021 DNS Challenge, which outperforms all top-performed models. We also carry out ablation experiments to tease apart all proposed sub-modules that are most important., Comment: Accepted by ICASSP 2022

Published
2021

17. Laparoscopic versus open surgery for perihilar cholangiocarcinoma: a multicenter propensity score analysis of short- term outcomes

Author

Wang, Min, Qin, Tingting, Zhang, Hang, Li, Jingdong, Deng, Xiaxing, Zhang, Yuhua, Zhao, Wenxing, Fan, Ying, Li, Dewei, Chen, Xuemin, Feng, Yechen, Zhu, Siwei, Xing, Zhongqiang, Yu, Guangsheng, Xu, Jian, Xie, Junjie, Dou, Changwei, Ma, Hongqin, Liu, Gangshan, Shao, Yue, Chen, Weibo, Liu, Jun, Liu, Jianhua, Yin, Xinmin, and Qin, Renyi

Published
2023
Full Text
View/download PDF

21. Effect of Laparoscopic and Open Pancreaticoduodenectomy for Pancreatic or Periampullary Tumors: Three-year Follow-up of a Randomized Clinical Trial

Author

Qin, Tingting, Zhang, Hang, Pan, Shutao, Liu, Jun, Li, Dewei, Chen, Rufu, Huang, Xiaobing, Liu, Yahui, Liu, Jianhua, Cheng, Wei, Chen, Xuemin, Zhao, Wenxing, Li, Jingdong, Tan, Zhijian, Huang, Heguang, Li, Deyu, Zhu, Feng, Yu, Guangsheng, Zhou, Baoyong, Zheng, Shangyou, Tang, Yichen, Ke, Jianji, Liu, Xueqing, Chen, Botao, Chen, Weibo, Ma, Hongqin, Xu, Jian, Liu, Yifeng, Lin, Ronggui, Dong, Yadong, Yu, Yahong, Wang, Min, and Qin, Renyi

Published
2024
Full Text
View/download PDF

22. NFAQP: Normalizing Flow Based Approximate Query Processing

Author

Cen, Libin, Li, Jingdong, Yue, Wenjing, Wang, Xiaoling, Goos, Gerhard, Founding Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Yang, Xiaochun, editor, Suhartanto, Heru, editor, Wang, Guoren, editor, Wang, Bin, editor, Jiang, Jing, editor, Li, Bing, editor, Zhu, Huaijie, editor, and Cui, Ningning, editor

Published
2023
Full Text
View/download PDF

30. The Long-Term Outcome of Laparoscopic Resection for Perihilar Cholangiocarcinoma Compared with the Open Approach: A Real-World Multicentric Analysis

Author

Qin, Tingting, Wang, Min, Zhang, Hang, Li, Jingdong, Deng, Xiaxing, Zhang, Yuhua, Zhao, Wenxing, Fan, Ying, Li, Dewei, Chen, Xuemin, Feng, Yechen, Zhu, Siwei, Xing, Zhongqiang, Yu, Guangsheng, Xu, Jian, Xie, Junjie, Dou, Changwei, Ma, Hongqin, Liu, Gangshan, Shao, Yue, Chen, Weibo, Xu, Simiao, Liu, Jun, Liu, Jianhua, Yin, Xinmin, and Qin, Renyi

Published
2023
Full Text
View/download PDF

31. Single Channel Speech Enhancement Using Temporal Convolutional Recurrent Neural Networks

Author

Li, Jingdong, Zhang, Hui, Zhang, Xueliang, and Li, Changliang

Subjects
Computer Science - Sound, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Audio and Speech Processing
Abstract

In recent decades, neural network based methods have significantly improved the performace of speech enhancement. Most of them estimate time-frequency (T-F) representation of target speech directly or indirectly, then resynthesize waveform using the estimated T-F representation. In this work, we proposed the temporal convolutional recurrent network (TCRN), an end-to-end model that directly map noisy waveform to clean waveform. The TCRN, which is combined convolution and recurrent neural network, is able to efficiently and effectively leverage short-term ang long-term information. Futuremore, we present the architecture that repeatedly downsample and upsample speech during forward propagation. We show that our model is able to improve the performance of model, compared with existing convolutional recurrent networks. Futuremore, We present several key techniques to stabilize the training process. The experimental results show that our model consistently outperforms existing speech enhancement approaches, in terms of speech intelligibility and quality.

Published
2020

33. Effect of Laparoscopic and Open Pancreatoduodenectomy for Pancreatic or Periampullary Tumors: Three-year Follow-up of a Randomized Clinical Trial

Author

Qin, Tingting, Zhang, Hang, Pan, Shutao, Liu, Jun, Li, Dewei, Chen, Rufu, Huang, Xiaobing, Liu, Yahui, Liu, Jianhua, Cheng, Wei, Chen, Xuemin, Zhao, Wenxing, Li, Jingdong, Tan, Zhijian, Huang, Heguang, Li, Deyu, Zhu, Feng, Yu, Guangsheng, Zhou, Baoyong, Zheng, Shangyou, Tang, Yichen, Ke, Jianji, Liu, Xueqing, Chen, Botao, Chen, Weibo, Ma, Hongqin, Xu, Jian, Liu, Yifeng, Lin, Ronggui, Dong, Yadong, Yu, Yahong, Wang, Min, and Qin, Renyi

Published
2023
Full Text
View/download PDF

34. Teacher-Student Training for Robust Tacotron-based TTS

Author

Liu, Rui, Sisman, Berrak, Li, Jingdong, Bao, Feilong, Gao, Guanglai, and Li, Haizhou

Subjects
Computer Science - Computation and Language, Computer Science - Sound, Electrical Engineering and Systems Science - Audio and Speech Processing
Abstract

While neural end-to-end text-to-speech (TTS) is superior to conventional statistical methods in many ways, the exposure bias problem in the autoregressive models remains an issue to be resolved. The exposure bias problem arises from the mismatch between the training and inference process, that results in unpredictable performance for out-of-domain test data at run-time. To overcome this, we propose a teacher-student training scheme for Tacotron-based TTS by introducing a distillation loss function in addition to the feature loss function. We first train a Tacotron2-based TTS model by always providing natural speech frames to the decoder, that serves as a teacher model. We then train another Tacotron2-based model as a student model, of which the decoder takes the predicted speech frames as input, similar to how the decoder works during run-time inference. With the distillation loss, the student model learns the output probabilities from the teacher model, that is called knowledge distillation. Experiments show that our proposed training scheme consistently improves the voice quality for out-of-domain test data both in Chinese and English systems., Comment: To appear at ICASSP2020, Barcelona, Spain

Published
2019

42. Establishment of a multifunctional impact system and a study of a pancreatic trauma model in rats based on controlling the injury area

Author

Wang Hailin, Han Li, Zhao Zhirong, Wang Qingqing, Li Jingdong, and Dai Ruiwu

Subjects
Impactor, Pancreatic trauma, Model, Injury area, Science (General), Q1-390, Social sciences (General), H1-99
Abstract

Background: At present, basic scientific research on pancreatic trauma is rare due to the lack of ideal animal models and modeling equipment for pancreatic trauma. Therefore, we intend to develop a multifunctional impact system with simple operation, diverse impact and accurate measurement and to establish a rat pancreatic trauma model based on injury area control by using the system. Methods: The impactor was designed based on the convenience of the impact energy acquisition, the diversity of the impact operation, and the precision of the impact strength parameter measurement by the team. The efficacy and stability/repeatability of the impactor were preliminarily evaluated. An impact head with different impact areas (3 cm2 and 6 cm2) of the impactor was used to squeeze the rat pancreas in the abdomen to form different injury areas under a pressure of 400 kPa. The efficacy features of this trauma model were evaluated by detecting the outcomes of pathology and biochemistry at 24 h after injury in the two groups. Furthermore, these changes were also evaluated at 6 h, 24 h, 48 h and 72 h after injury in the 3 cm2 trauma group. Result: Multifunctional impactors were successfully explored. The impact force was continuously adjustable with a range of 0–200 kg. The compression and extrusion stress ranges were continuously adjustable from 0 to 100 kg. System adjustment verified that the impactor had fine efficacy (P 0.05). Compared with the control group, rats in the pancreatic trauma group with different injury areas exhibited obvious injuries (P

Published
2023
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results