A5100682835- Robotics and Sensor-Based Localization
- Thermochemical Biomass Conversion Processes
- Advanced Vision and Imaging
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Image Processing Techniques and Applications
- Biofuel production and bioconversion
- Lignin and Wood Chemistry
- Image Enhancement Techniques
- Smart Grid and Power Systems
- Water Quality Monitoring Technologies
- Power Systems and Technologies
- Multimodal Machine Learning Applications
- Soft Robotics and Applications
- Robotic Path Planning Algorithms
- Biodiesel Production and Applications
- Reliability and Maintenance Optimization
- Geotechnical Engineering and Soil Mechanics
- Digital Imaging for Blood Diseases
- Structural Health Monitoring Techniques
- Robot Manipulation and Learning
- Energy Efficiency and Management
- Catalysis for Biomass Conversion
- Surgical Simulation and Training
Westlake University
2024
Second Hospital of Anhui Medical University
2024
Anhui Medical University
2024
Shanghai Jiao Tong University
2023-2024
Henan Normal University
2024
Beijing University of Posts and Telecommunications
2023
Space Engineering University
2023
Hebei University of Technology
2023
Curtin University
2020-2022
Beijing Satellite Navigation Center
2021-2022
Poly(ethylene oxide) (PEO)-based polymer electrolytes have shown extraordinary promise for all-solid-state lithium batteries; however, the practical application was severely restricted by their low ionic conductivity. In this work, robust pores of HKUST-1(Cu) were first filled a lithium-containing liquid (Li-IL) to form ion-conductive Li-IL@HKUST-1. Subsequently, flexible composite (CPEs) constructed via solution-casting approach upon incorporation Li-IL@HKUST-1 with PEO. The as-synthesized...
Abstract Fresh sweat contains a diverse range of physiological indicators that can effectively reflect changes in the body. However, existing wearable detection systems face challenges efficiently collecting and detecting fresh real‐time. Additionally, they often lack necessary deformation capabilities, resulting discomfort for wearer. Here, fully elastic electrochemical system is developed integrates sweat‐collecting microfluidic chip, multi‐parameter sensor, micro‐heater, circuit board...
This study proposes a control method that integrates deep reinforcement learning with load forecasting, to enhance the energy efficiency of ground source heat pump systems. Eight machine models are first developed predict future cooling loads, and optimal one is then incorporated into learning. Through interaction environment, strategy identified using Q-network optimize supply water temperature from source, allowing for savings. The obtained results show XGBoost model significantly...
A parallel fish school tracking based on multiple-feature detection has been proposed in this paper to obtain accurate movement trajectories of a large number zebrafish. Zebrafish are widely adapted many fields as an excellent model organism. Due the non-rigid body, similar appearance, rapid transition, and frequent occlusions, vision-based behavioral monitoring is still challenge. multiple appearance feature scheme was developed by examining head center body shape index features. The...
In this paper, we present a simple yet powerful deep neural network for natural image dehazing. The proposed method is designed based on U-Net architecture and made some design changes to make it better. We first use Group Normalization replace Batch solve the problem of insufficient batch size due hardware limitations. Second, introduce FReLU activation into block, which can achieve capturing complicated visual layouts with regular convolutions. Experimental results public benchmarks...
Captured underwater images usually suffer from severe color cast and low contrast due to wavelength-dependent light absorption scattering. These degradation issues affect the accuracy of target detection visual understanding. The image formation model is widely used improve quality images. Accurate transmission map background estimation are keys obtaining clear We develop a multi-priors restoration method with adaptive (MUAT). Concretely, we first propose calculation dominant channel cope...
High-pressure reactive distillation is a new bio-oil upgrading technology that can achieve high yields of the distillate due to reduced polymerization. Levoglucosan, as an anhydrous sugar, one major components from pyrolysis biomass. To understand at pressure, reactions and distribution levoglucosan during high-pressure were investigated in this study. Our results indicate would mainly undergo hydrolysis distillation, whereas thermal polymerization dominate atmospheric pressure distillation....
In this paper we proposed a method to implement estimation of mobile robot's position by using SURF (Speeded Up Robust Features) algorithm based on depth image in indoor environment. which is derived from SIFT has the advantage fast calculation speed and strong robustness. The generated 2D LRF (Laser Range Finder) sensor controlled rotate around its y-axis. According interest points each frame image, find relationship between two frames make match corresponding points. experiment, will show...
In past years, we have been dedicated to automating user acceptance testing (UAT) process of WeChat Pay, one the most influential mobile payment applications in China. A system titled XUAT has developed for this purpose. However, there is still a human-labor-intensive stage, i.e, test scripts generation, current system. Therefore, paper, concentrate on methods boosting automation level system, particularly stage generation. With recent notable successes, large language models (LLMs)...
Few-shot learning, especially few-shot image classification (FSIC), endeavors to recognize new categories using only a handful of labeled images by transferring knowledge from model trained on base categories. Despite numerous efforts address the challenge deficient transferability caused distribution shift between and classes, fundamental principles remain subject debate. In this paper, we elucidate why decline in performance occurs what information is transferred during testing phase,...
Detection of interaction force between soft tissue and surgical instrument perception biomechanical property are imperative for reducing operation time the risk surgeon mistake. However, in Robot-assisted minimally invasive laparoscopic surgery (MILS), surgeon’s ability perceiving valuable haptic information by manipulating robot system is severely impaired. In order to improve this robot, many university laboratories research institutions have been committing designing sensor developing...
Gradient-based meta-learning (GBML) algorithms are able to fast adapt new tasks by transferring the learned meta-knowledge, while assuming that all come from same distribution (in-distribution, ID). However, in real world, they often suffer an out-of-distribution (OOD) generalization problem, where different distributions. OOD exacerbates inconsistencies magnitudes and directions of task gradients, which brings challenges for GBML optimize meta-knowledge minimizing sum gradients each...
This paper presents a vision guided robot for recognizing and grasping moving objects on the conveyor belt. An industrial camera is utilized to capture image of objects. A robust system applied estimating object in real time. Objects are recognized by using SURF(Speeded-Up Robust Features) key point detector FLANN(Fast Library Approximate Nearest Neighbor) based matcher. Then Rotating calipers algorithm adopted estimate centroid position orientation simple path planning scheme will thus be...