A5100737916- Advanced Image and Video Retrieval Techniques
- Interconnection Networks and Systems
- Advanced Neural Network Applications
- Non-Destructive Testing Techniques
- Image Enhancement Techniques
- Anomaly Detection Techniques and Applications
- Optical Polarization and Ellipsometry
- Cell Image Analysis Techniques
- Photosynthetic Processes and Mechanisms
- Gait Recognition and Analysis
- Advanced Text Analysis Techniques
- Image and Signal Denoising Methods
- Parallel Computing and Optimization Techniques
- Network Traffic and Congestion Control
- Domain Adaptation and Few-Shot Learning
- Mechanics and Biomechanics Studies
- Magnetic Field Sensors Techniques
- Advanced Optical Imaging Technologies
- Plant responses to elevated CO2
- Distributed systems and fault tolerance
- Robotic Mechanisms and Dynamics
- Video Surveillance and Tracking Methods
- Plant Stress Responses and Tolerance
- Spectroscopy Techniques in Biomedical and Chemical Research
- Hand Gesture Recognition Systems
Southwest University of Science and Technology
2011-2023
China Jiliang University
2023
Nanjing Agricultural University
2023
Beijing University of Posts and Telecommunications
2022
Shenzhen University
2021
Hubei University of Technology
2021
Dalian University of Technology
2020
Nanjing University of Posts and Telecommunications
2018
This paper investigates a challenging problem,which is known as fine-grained image classification(FGIC). Different from conventional computer visionproblems, FGIC suffers the large intraclassdiversities and subtle inter-class differences.Existing approaches are limited to exploreonly visual information embedded in images.In this paper, we present novel approachwhich can use handy prior knowledge eitherstructured bases or unstructured text tofacilitate FGIC. Specifically, propose...
Image denoising is a fundamental problem in computer vision and multimedia computation. Non-local filters are effective for image denoising. But existing deep learning methods that use non-local computation structures mostly designed high-level tasks, global self-attention usually adopted. For the task of denoising, they have high computational complexity lot redundant uncorrelated pixels. To solve this combine marvelous advantages filter learning, we propose Convolutional Unbiased Regional...
Dempster-Shafer (D-S) evidence theory is regarded as an effective method of dealing with the uncertainty information but still suffers from conflict problem among evidence. Despite that various techniques were presented in plenty publications, currently, there no convincing and firm conclusion about optimum solution. This study presents Belief Coulomb Force (BCF) into D-S theory, where Zhou et al. entropy applied to represent electrical charge belief function, gravity or repulsion would be...
In this paper, an optical image encryption method is proposed based on structured light illumination and a diffractive neural network (DNN), which can realize conversion between different images. With the use of phase mask (SPM) in iterative retrieval algorithm, plaintext encoded into DNN composed multiple phase-only masks (POMs) ciphertext. It worth noting that ciphertext visible such one to another achieved, leading high concealment method. addition, wavelength illuminating light, all...
Abiotic stresses have had a substantial impact on fruit crop output and quality. Plants evolved an efficient immune system to combat abiotic stress, which employs reactive oxygen species (ROS) activate the downstream defence response signals. Although aquaporin protein encoded by PbPIP1;4 is identified from transcriptome analysis of Pyrus betulaefolia plants under drought treatments, little attention has been paid role PIP ROS in responding pear plants. In this study, we discovered that...
Object Detection Algorithms is widely used in transportation. With YOLOv3 however, it impossible to achieve real-time detection. This paper made some adjustments YOLOv3, and proposed a new light-scale model named MobileNetv1_yolov3lite. In our MobileNetv1_yolov3lite, we use MobileNetv1 instead of Darknet53 as backbone network, newly module yolov3lite for feature fusion. These significant increases detecting speed, can However, suffers from accuracy loss. order improve accuracy, further...
Magnetic sensor is a magnetic inductance sensor, which has the advantages of high sensitivity, wide measurement range, and low energy consumption compared to other field sensors, such as gate sensors sensors. are widely used in various domains metal detection, safety inspection, drone navigation. This paper describes ferromagnetic Object detection method based on sensor. uses obtain changes spatial intensity caused by continuous movement measured object over certain distance within thereby...
Thread level speculation (TLS) and Transactional memory (TM) are both promising way to enhance the performance of chip multiprocessor (CMP). The complexity providing efficient accesses buffering mechanism in TLS can be supported by TM logically. This paper proposes a speculative multi-threading model based on transactional memory, including its special hardware, compiler execution support. It's low-design-complexity approach effective unified support for TLS&TM. experimental results show...
The mechanical arm is a very complex system. It difficult to determine its initial position and motion trail due uncertainty perturbation of parameters external disturbance. In this paper, novel method determining the change in space proposed by using just movement such as coordinate value changes Cartesian system Spherical system, which different from other given previous literature. determined constructing an optimization objective function then calculating coordinates intelligent...
Abstract With the rapid development of IoT and image processing technology, there is an urgent need to solve problems clarity reliability recognition technology for monitoring systems in foggy environments. In this paper, a fast-dehazing algorithm based on complementary saturation prior proposed. First, mathematical model brightness estimate transmission established, which will simplify computation reduce running time program. Secondly, method proposed atmospheric light prior, find more...
This paper proposes a traffic police gesture recognition algorithm based on nine-axis Inertial measurement unit(IMU). Two wearable devices are attached to both hands of the tester, and then eight types standard gestures performed 150 times. During experiment, two-channel data three-axis acceleration, angular velocity, magnetometer from is collected for model training validation. A Butterworth filter utilized denoising Kalman fusion two channels. After that, feature extraction implemented...
Recently, Narrow- Band- Internet-of- Things are developing and being utilized in all industrial fields for the purpose of data transmission equipment monitoring. Due to existence middleware such as firewalls, TCP connection may be accidentally disconnected without a two-way heartbeat package. That means will not completely disconnected. In this paper, we have improved traditional packet sending algorithm by using an exponential linear algorithm, can set rate within reasonable intervals,...