Search

Your search keyword '"Liang, Mangui"' showing total 104 results
Start Over Author "Liang, Mangui"
104 results on '"Liang, Mangui"'

1. Soft-Weighted CrossEntropy Loss for Continous Alzheimer's Disease Detection

Author

Zhang, Xiaohui, Fu, Wenjie, and Liang, Mangui

Subjects
Computer Science - Sound, Electrical Engineering and Systems Science - Audio and Speech Processing, Quantitative Biology - Neurons and Cognition
Abstract

Alzheimer's disease is a common cognitive disorder in the elderly. Early and accurate diagnosis of Alzheimer's disease (AD) has a major impact on the progress of research on dementia. At present, researchers have used machine learning methods to detect Alzheimer's disease from the speech of participants. However, the recognition accuracy of current methods is unsatisfactory, and most of them focus on using low-dimensional handcrafted features to extract relevant information from audios. This paper proposes an Alzheimer's disease detection system based on the pre-trained framework Wav2vec 2.0 (Wav2vec2). In addition, by replacing the loss function with the Soft-Weighted CrossEntropy loss function, we achieved 85.45\% recognition accuracy on the same test dataset.

Published
2024

2. Multimodal Emotion Recognition from Raw Audio with Sinc-convolution

Author

Zhang, Xiaohui, Fu, Wenjie, and Liang, Mangui

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

Speech Emotion Recognition (SER) is still a complex task for computers with average recall rates usually about 70% on the most realistic datasets. Most SER systems use hand-crafted features extracted from audio signal such as energy, zero crossing rate, spectral information, prosodic, mel frequency cepstral coefficient (MFCC), and so on. More recently, using raw waveform for training neural network is becoming an emerging trend. This approach is advantageous as it eliminates the feature extraction pipeline. Learning from time-domain signal has shown good results for tasks such as speech recognition, speaker verification etc. In this paper, we utilize Sinc-convolution layer, which is an efficient architecture for preprocessing raw speech waveform for emotion recognition, to extract acoustic features from raw audio signals followed by a long short-term memory (LSTM). We also incorporate linguistic features and append a dialogical emotion decoding (DED) strategy. Our approach achieves a weighted accuracy of 85.1\% in four class emotion on the Interactive Emotional Dyadic Motion Capture (IEMOCAP) dataset.

Published
2024

3. TST: Time-Sparse Transducer for Automatic Speech Recognition

Author

Zhang, Xiaohui, Liang, Mangui, Tian, Zhengkun, Yi, Jiangyan, and Tao, Jianhua

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

End-to-end model, especially Recurrent Neural Network Transducer (RNN-T), has achieved great success in speech recognition. However, transducer requires a great memory footprint and computing time when processing a long decoding sequence. To solve this problem, we propose a model named time-sparse transducer, which introduces a time-sparse mechanism into transducer. In this mechanism, we obtain the intermediate representations by reducing the time resolution of the hidden states. Then the weighted average algorithm is used to combine these representations into sparse hidden states followed by the decoder. All the experiments are conducted on a Mandarin dataset AISHELL-1. Compared with RNN-T, the character error rate of the time-sparse transducer is close to RNN-T and the real-time factor is 50.00% of the original. By adjusting the time resolution, the time-sparse transducer can also reduce the real-time factor to 16.54% of the original at the expense of a 4.94% loss of precision., Comment: 10 pages

Published
2023

4. TST: Time-Sparse Transducer for Automatic Speech Recognition

Author

Zhang, Xiaohui, Liang, Mangui, Tian, Zhengkun, Yi, Jiangyan, Tao, Jianhua, 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, Fang, Lu, editor, Pei, Jian, editor, Zhai, Guangtao, editor, and Wang, Ruiping, editor

Published
2024
Full Text
View/download PDF

8. An Anonymous Identity-Based Authentication Scheme in Vector Network

Author

Yu, Jie, Liang, Mangui, Hutchison, David, Series editor, Kanade, Takeo, Series editor, Kittler, Josef, Series editor, Kleinberg, Jon M., Series editor, Mattern, Friedemann, Series editor, Mitchell, John C., Series editor, Naor, Moni, Series editor, Pandu Rangan, C., Series editor, Steffen, Bernhard, Series editor, Terzopoulos, Demetri, Series editor, Tygar, Doug, Series editor, Weikum, Gerhard, Series editor, Wang, Guojun, editor, Atiquzzaman, Mohammed, editor, Yan, Zheng, editor, and Choo, Kim-Kwang Raymond, editor

Published
2017
Full Text
View/download PDF

10. A Collaborative Reservation Mechanism of Multiple Parking Lots Based on Dynamic Vehicle Path Planning

Author

Xiao, Yuting, Cai, Zhen, Li, Jinglei, Liang, Mangui, and Long, Xiaoyu

Subjects
Telecommunication systems -- Usage, Reservation systems -- Usage, Parking lots -- Usage, Transportation industry
Abstract

With the development of wireless communication and artificial intelligence technology, online parking reservation system can effectively save drivers' searching time for vacant spaces. However, in the environment with multiple candidate parking lots around the destination, how to coordinate and maximize parking space resources to reduce the travel time is still a practical issue for urban drivers. In order to solve this problem, a collaborative reservation mechanism based on dynamic vehicle path planning is proposed in this paper. By the aid of the dedicated backbone network with a clear division of work responsibilities, the information of traffic and parking lots is collected in real time, based on which the travel time prediction and empty spaces evaluation are executed separately, and then the optimal decision of path planning and parking lot selection can be made and adjusted dynamically by a step-by-step acknowledgement mechanism. The simulation results show that, based on collaborative working and overall planning, our proposed reservation mechanism can effectively raise the utilization rate of the parking lots resources and significantly reduce the travel time for drivers under different traffic environments. Compared to current mechanisms, the collaborative parking reservation mechanism reveals higher feasibility and applicability. It can assist in design and operation of urban traffic management and space resource utilization., Author(s): Yuting Xiao [1]; Zhen Cai (corresponding author) [1,2]; Jinglei Li [1]; Mangui Liang [2]; Xiaoyu Long [3] 1. Introduction With the development of urbanization and the increase of population [...]

Published
2021
Full Text
View/download PDF

23. A Novel Addressing and Routing Architecture for Cloud-Service Datacenter Networks.

Author

Zhao, Aqun, Liu, Zhiyu, Pan, Jianping, and Liang, Mangui

Abstract

Datacenter networks (DCNs) play a key role in providing cloud services. The energy consumption and cost of a DCN are growing sharply with the extensions of network bandwidth and network size. The energy consumption, complexity and cost of a DCN depend on some design factors such as the topology structure, addressing scheme and routing mechanism. A novel addressing and routing architecture for cloud-service DCNs with regular topologies is proposed in this paper. First of all, we propose a port-based source-routing addressing (PSRA) scheme, which makes the table-lookup operation unnecessary and decreases the switch complexity. Next, leveraging the characteristics of PSRA and the regularity of DCN topologies, an extremely simple routing mechanism is designed, without switch involvement, control message interaction and topology information storage. Lastly, a high-efficiency fault-tolerance mechanism is proposed for the addressing and routing architecture. The analysis, implementation and simulation results indicate that the proposed architecture not only decreases the energy consumption and thus the cost of a DCN, but also enhances the routing performance and solves the fault-tolerance problem in a very efficient way. [ABSTRACT FROM AUTHOR]

Published
2022
Full Text
View/download PDF

35. Cepstral smoothing of masks for single-channel speech segregation

Author

Hu Qi, Liang Mangui, and Liao Niandong

Subjects
Noise, Voice activity detection, Computer Science::Sound, Computer science, Computational auditory scene analysis, Speech recognition, Cepstrum, Computer Science::Computation and Language (Computational Linguistics and Natural Language and Speech Processing), Speech processing, Linear predictive coding, Smoothing, Time–frequency analysis
Abstract

In this paper, cepstral smoothing is introduced to reduce musical noise in speech separated using computational auditory scene analysis (CASA). Our post-processing algorithm is composed of three steps. First, speech and interference are separated from noisy speech. Next, binary masks in the time-frequency domain are obtained using separated speech and noise signals. Finally, a mask smoothing algorithm which preserves speech onsets and quasi-stationary narrow-band structures is used to eliminate excessive peaks in time-frequency domain that lead to musical noise. Using an objective speech quality measure, we show that proposed method can improve speech quality.

Published
2010

37. A volume control unit based on TMS320C54

Author

Guo Yihe, Liang Mangui, Guo Yongfang, and Zhang Lu

Subjects
Audio signal, Adaptive algorithm, Microphone, Computer science, Noise (signal processing), Speech recognition, Audio signal flow, computer.software_genre, Signal, Adaptive filter, Signal-to-noise ratio, Electronic engineering, Audio signal processing, computer
Abstract

A unit controlling the volume of audio system in an automobile is designed. Main component in the unit is TMS320C54. We realize an adaptive algorithm with the chip to distinguish noise from outside and signal from the audio system. And the volume is controlled by signal to noise ratio (SNR). One microphone should be located somewhere in the automobile. The unit picks up the electrical signal from the microphone and electrical signals at all speakers in the audio system. Multi-channel signal are considered in the adaptive algorithm. We should decrease the complexity of the algorithm as low as possible with regard to our specific problem so that a TMS320C54 chip can implement the algorithm.

Published
2005

38. Design of the CTI system using SIP protocol

Author

Guo Yongfang, Liang Mangui, Guo Yihe, and Zhang Lu

Subjects
Service (systems architecture), Measure (data warehouse), Session Initiation Protocol, SIMPLE (military communications protocol), business.industry, Computer science, computer.internet_protocol, ComputerApplications_COMPUTERSINOTHERSYSTEMS, SIP trunking, Signaling protocol, Computer telephony integration, The Internet, business, computer, Computer network
Abstract

Using SIP (session initiation protocol), a CTI system is proposed for providing real-time and all-round service in call center. Compared with the traditional call center, for our CTI system adopts SIP as its signaling protocol, which can integrate other successful Internet service in a simple measure, such as HTTP and SMTP. The customer can be served through main kinds of convenient ways. Besides, call center becomes more flexible.

Published
2005

39. A design scheme of PSTN media gateway

Author

Guo Yongfang, Guo Yihe, Zhang Lu, and Liang Mangui

Subjects
Signaling protocol, Voice over IP, Computer science, business.industry, Default gateway, ComputerSystemsOrganization_COMPUTER-COMMUNICATIONNETWORKS, Next-generation network, business, Gateway address, Internetworking, H.248, Media gateway, Computer network
Abstract

This article presents an implementation of PSTN media gateway. The architecture of media gateway and the media processing scheme are described. It's designed from the idea of decomposed gateway. SIP, as a new VoIP technology, is adopted as the signaling protocol in the NGN. In the scheme we present the architecture of the hardware, the software control modules, and some other kernel technologies used in the media gateway.

Published
2005

43. High quality telephony over LAN

Author

Huang Guoming, Zhao Fanglong, Shen Ying, Wang Xiaoli, Shangguan Guanghua, Liang Mangui, and Ge Shelun

Subjects
Voice over IP, Transmission delay, business.industry, Computer science, Quality of service, ComputerSystemsOrganization_COMPUTER-COMMUNICATIONNETWORKS, Speech coding, Local area network, Network layer, Application layer, Telephony, business, Ethernet lan, Computer network
Abstract

This paper provides several research results about a technique of high quality telephony over LAN. Firstly, the value of the technique is discussed. Secondly, the delay performance of Ethernet LAN is tested by experiments both for the network layer and the application layer. Thirdly, a software architecture orientated to a real-time process like speech communication is proposed against the major reason of the overall transmission delay. Finally, as a conclusion, the possibility of realizing high quality telephony over LAN is demonstrated.

Published
2003

Catalog

Books, media, physical & digital resources
See catalog results