A5041145341- Metaheuristic Optimization Algorithms Research
- Video Surveillance and Tracking Methods
- Neural Networks and Applications
- Robotics and Sensor-Based Localization
- Advanced Multi-Objective Optimization Algorithms
- Face and Expression Recognition
- Robotic Path Planning Algorithms
- Chaos control and synchronization
- Adaptive Control of Nonlinear Systems
- Face recognition and analysis
- Evolutionary Algorithms and Applications
- Social Robot Interaction and HRI
- Industrial Technology and Control Systems
- Human Pose and Action Recognition
- EEG and Brain-Computer Interfaces
- Advanced Image and Video Retrieval Techniques
- Electric Vehicles and Infrastructure
- Machine Learning and ELM
- Anomaly Detection Techniques and Applications
- Advanced Neural Network Applications
- Multimodal Machine Learning Applications
- Robotics and Automated Systems
- Advanced Algorithms and Applications
- Hand Gesture Recognition Systems
- Advanced Database Systems and Queries
Guizhou University
2016-2025
Oklahoma State University
2017
Chinese Academy of Sciences
2011
Huazhong University of Science and Technology
2011
Hunan Institute of Science and Technology
2006
Prognostics, such as remaining useful life (RUL) prediction, is a crucial task in condition-based maintenance. A major challenge data-driven prognostics the difficulty of obtaining sufficient number samples failure progression. However, for traditional machine learning methods and deep neural networks, enough training data prerequisite to train good prediction models. In this work, we proposed transfer algorithm based on Bi-directional Long Short-Term Memory (BLSTM) recurrent networks RUL...
This study proposes a modified convolutional neural network (CNN) algorithm that is based on dropout and the stochastic gradient descent (SGD) optimizer (MCNN-DS), after analyzing problems of CNNs in extracting convolution features, to improve feature recognition rate reduce time-cost CNNs. The MCNN-DS has quadratic CNN structure adopts rectified linear unit as activation function avoid problem accelerate convergence. To address overfitting problem, uses an SGD optimizer, which implemented...
We adopted actual intelligent production requirements and proposed a tiny part defect detection method to obtain stable accurate real-time system solve the problems of manually setting conveyor speed industrial camera parameters in for factory products. First, we considered important influences properties parts environmental on its stability. Second, established correlation model between capability coefficient moving conveyor. Third, algorithm that are based single short detector network...
In order to realize fast real-time positioning after a mobile robot starts, this paper proposes an improved ORB-SLAM2 algorithm. Firstly, we proposed binary vocabulary storage method and training algorithm based on Oriented FAST Rotated BRIEF (ORB) operator reduce the size improve loading speed of tracking accuracy. Secondly, offline map construction element keyframe database; then, designed reposition map. Finally, presented visualization for elements mapping trajectories. check performance...
Many text mining tasks such as retrieval, summarization, and comparisons depend on the extraction of representative keywords from main text. Most existing keyword algorithms are based discrete bag-of-words type word representation In this paper, we propose a patent algorithm (PKEA) distributed Skip-gram model for classification. We also develop set quantitative performance measures evaluation information gain cross-validation, Support Vector Machine (SVM) classification, which valuable when...
Abstract Non‐suicide self‐injury (NSSI) can be dangerous and difficult for guardians or caregivers to detect in time. NSSI refers when people hurt themselves even though they have no wish cause critical long‐lasting hurt. To timely identify effectively prevent order reduce the suicide rates of patients with a potential risk, detection based on spatiotemporal features indoor activities is proposed. Firstly, an behaviour dataset provided, it includes four categories that used scientific...
Abstract Keyframe extraction can effectively help users quickly understand video content. Generally, keyframes should be representative of the content and simultaneously diverse to reduce redundancy. Aiming find features frames filter out video, we propose a method keyframe recommendation based on feature intercross fusion (KFRFIF). The is inspired by implied relations between keyframe-extraction problem problem. First, investigate application framework Second, architecture proposed KFRFIF...
As the problem of an aging population becomes more and serious, social robots have increasingly significant influence on human life. By employing regular question-and-answer conversations or wearable devices, some robotics products can establish personal health archives. But those are unable to collect diet information automatically through robot vision audition. A healthy reduce a person's risk developing cancer, diabetes, heart disease, other age-related diseases. In order perceive dietary...
Audio magnetotelluric (AMT) is commonly used in mineral resource exploration. However, the weak energy of AMT signals makes them susceptible to being overwhelmed by noise, leading erroneous geophysical interpretations. In recent years, deep learning has been applied denoising and shown better performance compared traditional methods. current methods overlook characteristics signals, resulting reduced accuracy. To enhance matching features we propose a CBAM-based (Convolutional Block...
Abstract To suppress the non‐linear motion for a class of strict‐feedback fractional order systems, and to improve their transient steady‐state performance, neuro‐adaptive prescribed performance backstepping control strategy suitable systems is proposed in this paper. Firstly, interval Type‐2 fuzzy neural network constructed approximate unknown functions. Secondly, tracking differentiator introduced address problem ‘explosion complexity’ associated with technique framework backstepping....
Multi-object tracking (MOT) tasks face challenges from multiple perception views due to the diversity of application scenarios. Different (front-view and top-view) have different imaging data distribution characteristics, but current MOT methods do not consider these differences only adopt a unified association strategy deal with various occlusion situations. This paper proposed View Adaptive Multi-Object Tracking Method Based on Depth Relationship Cues (ViewTrack) enable adapt scene's...
Various studies have focused on feature extraction methods for automatic patent classification in recent years. However, most of these approaches are based the knowledge from experts related domains. Here we propose a hierarchical model (HFEM) multi-label mechanical classification, which is able to capture both local features phrases as well global and temporal semantics. First, n-gram extractor convolutional neural networks (CNNs) designed extract salient lexical-level features. Next, long...