A5112289284- Viral Infections and Vectors
- Mosquito-borne diseases and control
- Domain Adaptation and Few-Shot Learning
- Advanced Neural Network Applications
- Vector-borne infectious diseases
- Advanced Measurement and Metrology Techniques
- Cell Image Analysis Techniques
- Optimal Experimental Design Methods
- Robotic Locomotion and Control
- Gait Recognition and Analysis
- Educational Reforms and Innovations
- Advanced Fluorescence Microscopy Techniques
- Anomaly Detection Techniques and Applications
- Robot Manipulation and Learning
- Characterization and Applications of Magnetic Nanoparticles
- Advanced Vision and Imaging
- Multimodal Machine Learning Applications
- Interactive and Immersive Displays
- Biometric Identification and Security
- Strategic Planning and Analysis
- Single-cell and spatial transcriptomics
- Natural Compounds in Disease Treatment
- Image Processing and 3D Reconstruction
- Teleoperation and Haptic Systems
- Advanced Multi-Objective Optimization Algorithms
University of Electronic Science and Technology of China
2024-2025
Chongqing University
2024
Wuhan University of Technology
2024
Hong Kong Polytechnic University
2024
University Town of Shenzhen
2024
Tsinghua University
2024
Amazon (Germany)
2023
Xi’an University of Posts and Telecommunications
2023
University of Chinese Academy of Sciences
2020
Albany Research Institute
2016
The recovery rate of magnetic beads is a crucial factor in evaluating the performance flow fluorescence immunoassay analyzer. This paper proposes an optimal design method for analyzer using orthogonal testing and comprehensive sensitivity analysis. experimental was employed to reduce number analysis tests required key parameters Through rigorous evaluation, combination adsorption height, liquid volume, time determined. results demonstrate that this can significantly improve beads. provides...
Recently, deep convolutional neural networks (DCNN) have been widely used in semantic segmentation tasks and achieved high accuracy. However, most algorithms based on DCNN computational complexity, making them unsuitable for real‐time segmentation. To solve this problem, paper proposes a algorithm the STDC network. The adopts an “encoder–decoder” embedded U‐shaped architecture to realize while maintaining Following encoder, mixed pooling attention module is designed expand receptive field,...
Low quality capture and obstruction on fingers often result in partially visible fingerprint images, which imposes challenge for recognition. In this work, motivated from the practical use cases, we first systematically studied different types of partial occlusion. Specifically, two major occlusion, including six granular types, corresponding methods to simulate each type model evaluation improvement were introduced. Second, proposed a novel Robust Partial Fingerprint (RPF) recognition...
This study adopts SWOT analysis to analyze Wuhan's talent policy in depth, aiming explore its advantages, disadvantages, opportunities and challenges. The finds that Wuhan has significant geographical location resource advantages terms of policy, as well outstanding performance innovation cultivation. However, there are also problems such drain insufficient implementation. In the face fierce competition for talents at home abroad, should seize opportunity strengthen promotion implementation,...
Instruction detection technology is a new generation of security that monitor networks or systems to avoid malicious activity and policy violation. Compared with traditional protection measures such as firewall, instruction can prevent attacks both from external internal. The SVM statistical learning model(SLT), which shows an extraordinary advantage when dealing small sample. Its advantages are: (1) SVM's goal get the optimal solution under limited samples but not infinity prerequisite...
Due to circuit failures, defective elements that cannot adaptively adjust the phase shifts of their impinging signals in a desired manner may exist on an intelligent reflecting surface (IRS). Traditional way find these IRS requires thorough diagnosis all circuits belonging huge number elements, which is practically challenging. In this paper, we will devise novel approach under transmitter sends known pilot and receiver localizes just based its over-the-air measurements reflected from IRS....
While tangible user interface has shown its power in naturally interacting with rigid or soft objects, users cannot conveniently use different types of granular materials as the interaction media. We introduce DipMe a smart device to recognize media real time, which can be used connect physical world various virtual content. Other than vision-based solutions, we propose dip operation our and exploit haptic signals materials. With modern machine learning tools, find from are distinguishable...
Deeplab series semantic segmentation algorithms extract target features using deep layers of a convolutional neural network, resulting in lacking detailed information, such as edges and shapes extracted by shallow layers. Deeplabv3plus uses atrous convolution to obtain feature maps, which lose some image information. All the above have an impact on performance improvement. In response these issues, reduce performance, we propose algorithm based multi-expert system that builds multiple expert...
In recent years, semantic segmentation methods based on deep learning have made remarkable developments. Despite achieving high accuracy, the performance of real-time cannot satisfy real-world applications. order to achieve a balance between accuracy and speed, algorithm Tversky loss function mixed pooling is proposed in this paper. A Short-Term Dense Concatenate network (STDC network) used construct encoder, module for final part which uses strip average enhance feature representation while...
In this paper, we propose a technique for making humans in photographs protrude like reliefs. Unlike previous methods which mostly focus on the face and head, our method aims to generate art works that describe whole body activity of character. One challenge is there no ground-truth supervised deep learning. We introduce sigmoid variant function manipulate gradients tactfully train neural networks by equipping with loss defined gradient domain. The second actual often across different light...