- Speech and Audio Processing
- Advanced Data Compression Techniques
- Advanced Algorithms and Applications
- Speech Recognition and Synthesis
- Advanced Multi-Objective Optimization Algorithms
- Metaheuristic Optimization Algorithms Research
- Neural dynamics and brain function
- Hearing Loss and Rehabilitation
- Advanced Computational Techniques and Applications
- Digital Filter Design and Implementation
- Acoustic Wave Phenomena Research
- Advanced Data Storage Technologies
- Indoor and Outdoor Localization Technologies
- Evolutionary Algorithms and Applications
- Advanced Measurement and Detection Methods
- Grey System Theory Applications
- Advanced Adaptive Filtering Techniques
- Computational Physics and Python Applications
- Caching and Content Delivery
- Image and Signal Denoising Methods
- Electricity Theft Detection Techniques
- Domain Adaptation and Few-Shot Learning
- Noise Effects and Management
- Image and Video Quality Assessment
- Neuroscience and Music Perception
Huashan Hospital
2020-2025
Fudan University
2020-2025
China Southern Power Grid (China)
2021-2024
State Key Laboratory of Medical Neurobiology
2021
Georgia Institute of Technology
2021
Hebei University of Technology
2006-2020
PLA Army Engineering University
2000-2019
Academy of Military Transportation
2018-2019
Anhui Science and Technology University
2019
East China Normal University
2018
Increasing absorption efficiency and decreasing total thickness of the acoustic absorber is favorable to promote its practical application. Four compressed porous metals with compression ratios 0%, 30%, 60%, 90% were prepared assemble four-layer gradient metals, which aimed develop high-efficiency thin thickness. Through deriving structural parameters thickness, porosity, static flow resistivity for theoretical models sound coefficients constructed through transfer matrix method according...
One of the most important benefits virtualization is Virtual Machine (VM) migration. While performance live VM migration well explored, security aspects have received very little attention. The extension trusted computing to virtual systems using vTPMs allows applications in use vTPM for secure storage and reporting platform integrity. In this paper, we propose an improved protocol a channel property-based attestation destination assure requirements
Objective: Vascular cognitive impairment (VCI) is a common complication in adult patients with moyamoya disease (MMD), and reversible by surgical revascularization its early stage of mild VCI. However, accurate diagnosis VCI difficult based on neuropsychological examination alone. This study proposed method dynamic resting-state functional connectivity (FC) network to recognize global MMD. Methods: For MMD, 36 43 intact cognition (Non-VCI) were included, as well 26 normal controls (NCs)....
Recently, wireless local area networks (WLANs) have become ubiquitous. In high density WLANs, interference resulting from neighbor cells is a challenging problem which would affect the network performance. To get higher throughput, it important to improve spatial reuse of channels. Since transmit power control (TPC) may cause asymmetric links unless carefully designed, we demonstrate that dynamic sensitive (DSC) can be used capacity in dense environment. order reduce influence inter-cell and...
When unmanned aerial vehicles (UAVs) support the Corps of Engineers in reconnaissance operations, order to gather visible image information that should meet mission’s need, we grouped engineering interpretation tasks into 10 levels by using National Imagery Interpretability Rating Scale (NIIRS). The quantitative relationship between targets, sensor performance, and flight altitude was established through general quality equation (GIQE) geometrical property ground sampled distance (GSD)....
This paper presents a spectral enhancement method to improve the quality of speech reconstructed by neural waveform generators with low-bit quantization. At training stage, this builds multiple-target DNN, which predicts log amplitude spectra natural high-bit waveforms together ratios between and distorted spectra. Log are adopted as model input. generation enhanced obtained an ensemble decoding strategy, further combined phase produce final inverse STFT. In our experiments on WaveRNN...
This paper presents the design methodology and implementation of reconfigurable coordinate rotation digital computer (CORDIC) architecture that can be configured to operate in different modes rotations achieve singleprecision floating point division, multiplication square-root operations. Through introducing pre- post-processing, float-point operations integrated into a unified CORDIC iteration procedure. According characteristics operations, we propose pipeline-parallel mixed optimize...
In order to attract more customers, it is a common practice for retailers provide advance sales, example, Maxim’s Bakery in Hong Kong, Amazon.com, Movies Unlimited and Toys R Us. Similarly, suppliers often allow their permissible delay payment increase sales. Advance sales trade credit policies numerous benefits companies, including gaining additional discriminative customers increased profit due interest earned from payments received committed prior the start of regular...
The competitive relationships among ports become complicated as the consequence of prosperity international trade. Typical oligopoly competition models cannot be competent for analysis in real scenario. In this paper, scale-free network is adopted to characterize interactions with various number neighbors. For each port node, not only direct but also indirect influences exerted by its neighbor are taken into account. Following hypothesis evolutionary game theory, social learning behavior...
In order to obtain some head-related impulse responses (HRIRs) not measured in virtual sound system, a real-time interpolation method based on the minimum-phase reconstruction is proposed this article. The composed of three steps: firstly, original HRIRs are preprocessed and separated into two parts: version corresponding inter-aural time difference (ITD); secondly, each part interpolated by using simple linear interpolation; finally, HRIR obtained from above parts. simulation experiments...
The simplicity and success of cuckoo search (CS) algorithm has inspired researchers to apply these techniques the multi-objective optimization field. paper studies application CS for solving problems (MOPs) based on decomposition methods. A new decomposition-based is proposed, called MOCS/D. proposed integrates unique Lévy flights technique improved polynomial mutation into evolutionary Decomposition (MOEA/D). Our approach compared with MOEA/D-SBX MOEA/D-DE test instances. experimental...
This study shows how the performance of a parallel simulation may be affected by structure system being simulated. We consider wide class “linearly synchronous” simulations consisting asynchronous and synchronous (or other distributed‐processing systems), with conservative or optimistic protocols, in which differences virtual times logical processes simulated real time t are order o ( ) as tends to infinity. Using random transformation idea, we show simulation′s processing rate is related...
The current method of smart meter verification relies on manual regular sampling inspection, which is heavy in workload and poor real-time, can’t fully monitor all the equipments. Therefore, a remote real-time error monitoring algorithm indispensable. We propose estimation model based genetic optimized Levenberg-Marquarelt (LM) algorithm. Firstly, law conservation energy, relationship between electricity consumption established. Then, LM used to estimate operating meter. Finally, we actual...
In city downtown, parking problem are becoming more and crucial due to limited spaces. This paper introduces a robust vehicle detection algorithm in lot based on wireless sensor network (WSN). the algorithm, three-axis anisotropic magneto-resistive (AMR) is used. The proposed tested lot. test results suggest that can effectively detect with an accuracy of 99%.
The applications of speech fractal dimension to segmentation and enhancement are researched. computer simulation results show that the new methods can achieve satisfactory performances.
Adversarial regularization has been shown to improve the generalization performance of deep learning models in various natural language processing tasks. Existing works usually formulate method as a zero-sum game, which is solved by alternating gradient descent/ascent algorithms. Such formulation treats adversarial and defending players equally, undesirable because only player contributes performance. To address this issue, we propose Stackelberg Regularization (SALT), formulates game. This...
Abstract In natural conditions, human and animals need to extract target sound information from noisy acoustic environments for communication survival. However, how the contextual environmental sounds impact tuning of central auditory neurons source azimuth over a wide range levels is not fully understood. Here, we determined azimuth‐level response areas ( ALRA s) rat cortex by recording their responses probe tones varying with azimuths under both quiet (probe alone) forward masking...