A5100691354- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Robotics and Sensor-Based Localization
- Visual Attention and Saliency Detection
- Image and Object Detection Techniques
- Neuroscience and Neural Engineering
- 3D Surveying and Cultural Heritage
- EEG and Brain-Computer Interfaces
- Video Surveillance and Tracking Methods
- Augmented Reality Applications
- Advanced Measurement and Detection Methods
- Gaze Tracking and Assistive Technology
- Magnetic Bearings and Levitation Dynamics
- Face and Expression Recognition
- Advanced Neural Network Applications
- Video Analysis and Summarization
- Medical Image Segmentation Techniques
- Advanced Wireless Network Optimization
- Advanced Vision and Imaging
- Face recognition and analysis
- Multimodal Machine Learning Applications
- Optical measurement and interference techniques
- Water Quality Monitoring Technologies
- Generative Adversarial Networks and Image Synthesis
- Cooperative Communication and Network Coding
Power Grid Corporation (India)
2025
Shanghai Jiao Tong University
2010-2024
Tianjin University of Technology and Education
2024
Powerchina Huadong Engineering Corporation (China)
2024
Southwest Petroleum University
2023-2024
Hohai University
2024
Minjiang University
2024
Beijing Institute of Technology
2015-2022
Southwest Jiaotong University
2004-2020
Chongqing University
2011-2017
Objective.The electroencephalography (EEG)-based brain-computer interfaces (BCIs) have been used in the control of robotic arms. The performance non-invasive BCIs may not be satisfactory due to poor quality EEG signals, so shared strategies were tried as an alternative solution. However, most existing methods set arbitration rules manually, which highly depended on specific tasks and developer's experience. In this study, we proposed a novel model that automatically optimized commands...
It is ambitious to develop a brain-controlled robotic arm for some patients with motor impairments perform activities of daily living using brain–computer interfaces (BCIs). Despite much progress achieved, this mission still very challenging mainly due the poor decoding performance BCIs. The problem even exacerbated in case noninvasive A shared control strategy developed work realize flexible reach and grasp multiple objects. With intelligent assistance provided by robot vision,...
Due to the combined advantages of no-contact friction and self-stable levitation, high temperature superconducting magnetic levitation (HTS Maglev) has significant potential for rail transit applications. In order further improve carrying capacity HTS maglev system, it is necessary optimize permanent magnet guideway (PMG). this paper, original Halbach PMG was optimized a new with better performance designed manufactured. The field forces were calculated by finite element method, size...
Deep learning-based methods have shown excellent potential on object detection and pose estimation with vast amounts of training data to achieve good performance. Obtaining enough comprehensive manual labeling is a time-consuming error-prone task in industrial scenes, most current time-saving synthetic generation require detailed high-quality CAD models. Instead, we introduce an image-to-image translation based approach, which requires only untextured models small number real images. The...
Stereo-electroencephalographic (SEEG) depth electrodes were used to record neural activity from deep brain structures in this study. By localizing all the into individual brain, we found that areas are inside of central sulcus occurred obvious hand-movement-related modulation when subjects performing different hand motion tasks. Then, an asynchronous brain-computer interface which enables subject control a prosthetic real time was built. The testing results showed that, using SEEG signals...
Objective. White matter tissue takes up approximately 50% of the human brain volume and it is widely known as a messenger conducting information between areas central nervous system. However, characteristics white neural activity whether recordings can contribute to movement decoding are often ignored still remain largely unknown. In this work, we make quantitative analyses investigate these two important questions using invasive recordings. Approach. We recorded...
Myoelectric control of multifunctional prostheses is challenging for individuals with high-level amputations due to insufficient surface electromyography (sEMG) signals. A surgical technique called targeted muscle reinnervation (TMR) has achieved impressive improvements in myoelectric by providing more sEMG In this case, some channels signals are coupled after TMR, which limits the performance conventional amplitude-based upper-limb prostheses. paper, two different ways (training and...
Reinforcement learning (RL) is pivotal in empowering Unmanned Aerial Vehicles (UAVs) to navigate and make decisions efficiently intelligently within complex dynamic surroundings. Despite its significance, RL hampered by inherent limitations such as low sample efficiency, restricted generalization capabilities, a heavy reliance on the intricacies of reward function design. These challenges often render single-method approaches inadequate, particularly context UAV operations where high costs...
This article proposed a new design of dc linear motor for the propulsion high-temperature superconducting (HTS) magnet levitation (maglev) transports. is effective because it can reduce HTS maglev system cost and compact components by improving integration three main functions levitation, guidance, propulsion. The was made desired onboard coil so that applicable to existing track system. Meanwhile, previous independent be replaced. idea first evaluated finite-element simulation optimized...
An important step in content-based video retrieval is the temporal segmentation of video. In this paper we give an overview some existing shot boundary detection algorithms. that operate on uncompressed stream. After that, propose algorithm for integrates spatial and color features frames. Our method not sensitive to brightness change quick motion. Therefore, it can improve precision detecting boundaries. Finally experimental results draw our conclusion.
Convolutional neural networks have shown excellent potential on establishing correspondences from 2D images to 3D objects for object 6D pose estimation, both dense and sparse methods. However, only single geometric representation between each pixel keypoint is utilized in existing In this work, we attempt explore more accurate predictions with multiple representations the method. First, utilize convolutional network regress pixel-wise offset vector field, convert field into directions...
Industrial augmented reality (AR) applications demand high on the visual consistency of virtual-real registration. To present, marker-based registration method is most popular because it fast, robust, and convenient to obtain matrix. In practice, matrix should multiply an offset that describes transformation between attaching position initial marker relative object. However, usually measured, calculated, set manually, which not accurate convenient. This paper proposes automatic marker–object...
Channel correlation between users can significantly affect the system performance in multiuser multiple-input multiple-output (MU-MIMO) systems. To prevent highly correlated from being scheduled same radio blocks (RBs), these are separated into different groups. Most of related papers focus on how to separate according their channel correlations. However, group number also plays an important role and this is not investigated those papers. Therefore, paper we investigate MU-MIMO systems...