A5091623026- Industrial Vision Systems and Defect Detection
- Digital Media Forensic Detection
- Video Surveillance and Tracking Methods
- Advanced Steganography and Watermarking Techniques
- Advanced Algorithms and Applications
- Infrared Target Detection Methodologies
- Advanced Neural Network Applications
- Advanced Computational Techniques and Applications
- Chaos-based Image/Signal Encryption
- Face and Expression Recognition
- Image and Object Detection Techniques
- Color Science and Applications
- Image Processing Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- Vehicle License Plate Recognition
- Image and Video Stabilization
- Advanced Measurement and Detection Methods
- Advanced Image and Video Retrieval Techniques
- Image Enhancement Techniques
- Human Pose and Action Recognition
- Biometric Identification and Security
- Cell Image Analysis Techniques
- Automated Road and Building Extraction
- Face recognition and analysis
- Industrial Technology and Control Systems
China West Normal University
2006-2024
Xi'an University of Technology
2012-2024
Xidian University
2023
Nanjing University of Posts and Telecommunications
2022
Inner Mongolia Electric Power Survey & Design Institute (China)
2021
Baotou Medical College
2021
China Electronics Standardization Institute
2021
Shanxi Coal Transportation and Sales Group (China)
2020
Shaoguan Railway Hospital
2009-2019
Kunming University of Science and Technology
2018
Due to the limited computing resources and storage capacity of edge detection devices, efficient algorithms are typically required meet real-time accuracy requirements. Existing detectors often require a large number parameters high computational power improve accuracy, which reduces speed performance on low-power devices. To reduce load enhance we propose lightweight drone target algorithm, LD-YOLOv10. Firstly, design novel feature extraction structure called RGELAN, utilizes...
Detection of weeds and crops is the key step for precision spraying using herbicide robot precise fertilization agriculture machine in field. On basis k-mean clustering image segmentation color information connected region analysis, a method combining multi feature fusion support vector (SVM) was proposed to identify detect position corn seedlings weeds, reduce harm on growth, achieve accurate fertilization, thereby realizing weeding or fertilizing. First, dataset weed seedling...
Glucose-6-phosphate dehydrogenase (G6PD) deficiency is one of the most common X-linked enzymopathies caused by G6PD gene variant. We aimed to provide characteristics and variant distribution in a large Chinese newborn screening population. investigated prevalence China from 2013 2017. Then, we examined activity representative birth cohort explore 2016. then performed multicolor melting curve analysis classify variants 10,357 neonates with activity-confirmed deficiency, DNA Sanger sequencing...
Automatic age classification has drawn significant interest in plenty of applications such as access control, human-computer interaction, law enforcement and surveillance. is a challenging task due to the complexity facial images. A large number approaches have been investigated on unconstrained datasets. However, most these focused network architecture rather than distribution data, i.e., extreme class imbalance existing among different groups difficulty data collection. In this paper, we...
In the realm of traffic sign detection, challenges arise due to small size objects, complex scenes, varying scales signs, and dispersed objects. To address these problems, this paper proposes a object detection algorithm, YOLOv8s-DDA, for signs based on an improved YOLOv8s. Specifically, C2f-DWR-DRB module is introduced, which utilizes efficient two-step method capture multi-scale contextual information employs dilated re-parameterization block enhance feature extraction quality while...
In the printing industry, defect detection is of crucial importance for ensuring quality printed matter. However, rarely has research been conducted web offset printing. this paper, we propose an automatic method printing, which consists determining first row captured images, image registration and detection. Determining images a particular problem not studied before. To solve problem, fast computational algorithm based on projection given, can convert 2D searching into 1D feature matching....
With a good balance between tracking accuracy and speed, kernelized correlation filter (KCF)-based trackers have now become dominant approaches in the field of online object tracking. However, KCF-based tracker updates templates by linearly combining historical objects, its performance inevitably degrades when accumulated error causes model to look more similar background than foreground. To address this drifting problem, we propose an incremental kernel principal component analysis...
Deep learning has been successfully applied to classification tasks in many fields due its good performance discriminative features. However, the application of deep printing defect is very rare, and there almost no research on method for defects with imbalanced samples. In this paper, we present a convolutional neural network model extract features directly from printed image defects. Furthermore, considering asymmetry number different types samples—that is, kinds samples unbalanced—seven...
In this paper, a novel method using wavelet decomposition and 2D Principal component analysis (2DPCA) for palmprint recognition is presented. Firstly, transform adopted to obtain different level of coefficients the original image; secondly 2DPCA applied on low-frequency that contains most discrimination information image. One criterion not all PCs are useful demonstrated rule selecting proposed. Lastly, algorithm tested PolyU image database experimental result encouraging achieves...
Diameter detection plays an important role in the production of circular workpieces. It is very difficult to obtain accurate measurement diameter when workpiece red-hot because existence lots noise. Although Hough transform (HT) has many advantages for workpieces, it still suffered from some serious problems, which include requirements a large amount computation and memory occupancy. To address these we propose intelligent method with Jaya (IHTJ). The key points are as follows. We analyze...
Different devices in the smart home environment are subject to different levels of attack. Devices with lower attack frequencies confront difficulties collecting data, which restricts ability train intrusion detection models. Therefore, this paper presents a novel method called EM-FEDE (enhancement based on feature enhancement and data enhancement) generate adequate training for expanding few-shot datasets. Training models an expanded dataset can enhance performance. Firstly, adaptively...