A5100364049- Video Surveillance and Tracking Methods
- Advanced Vision and Imaging
- Human Pose and Action Recognition
- Advanced Image and Video Retrieval Techniques
- Visual Attention and Saliency Detection
- Image Enhancement Techniques
- Fire Detection and Safety Systems
- Infrared Target Detection Methodologies
- Evolutionary Game Theory and Cooperation
- Video Analysis and Summarization
- Experimental Behavioral Economics Studies
- Evolutionary Psychology and Human Behavior
- Advanced Image Processing Techniques
- Video Coding and Compression Technologies
- Image and Video Quality Assessment
- Face recognition and analysis
- Opinion Dynamics and Social Influence
- Remote Sensing and Land Use
- Anomaly Detection Techniques and Applications
- Mathematical and Theoretical Epidemiology and Ecology Models
- Remote-Sensing Image Classification
- Multimodal Machine Learning Applications
- Music and Audio Processing
- AI and Multimedia in Education
- Face and Expression Recognition
Northwestern Polytechnical University
2015-2024
Institute of Intelligent Machines
2018
Chinese Academy of Sciences
2018
University of Science and Technology of China
2010-2018
Northwest Normal University
2018
Dahua Technology (China)
2018
Huawei Technologies (China)
2013
PLA Army Engineering University
2011
Nanyang Technological University
2011
Unsupervised video object segmentation aims to automatically segment moving objects over an unconstrained without any user annotation. So far, only few unsupervised online methods have been reported in the literature, and their performance is still far from satisfactory because complementary information future frames cannot be processed under setting. To solve this challenging problem, paper, we propose a novel (UOVOS) framework by construing motion property mean concurrence with generic for...
Human action analysis and understanding in videos is an important challenging task. Although substantial progress has been made past years, the explainability of existing methods still limited. In this work, we propose a novel reasoning framework that uses prior knowledge to explain semantic-level observations video state changes. Our method takes advantage both classical modern deep learning approaches. Specifically, defined as information target domain, including set objects, attributes...
The target representation learned by convolutional neural networks plays an important role in Thermal Infrared (TIR) tracking. Currently, most of the top-performing TIR trackers are still employing representations model trained on RGB data. However, this does not take into account information modality itself, limiting performance
Tracking-by-learning strategies have been effective in solving many challenging problems visual tracking, which the learning sample generation and labeling play important roles for final performance. Since concern of deep based approaches has shown an impressive performance different vision tasks, how to properly apply model, such as CNN, online tracking framework is still challenging. In this paper, overcome overfitting problem caused by straight-forward incorporation, we propose...
In urban environment monitoring, visual tracking on unmanned aerial vehicles (UAVs) can produce more applications owing to the inherent advantages, but it also brings new challenges for existing approaches (such as complex background clutters, rotation, fast motion, small objects, and realtime issues due camera motion viewpoint changes). Based Siamese network, be conducted efficiently in recent UAV datasets. Unfortunately, learned convolutional neural network (CNN) features are not...
Abstract Correlation filters with convolutional neural network (CNN) features have been successfully applied to visual tracking owing their impressive combined capability for object representation. Unfortunately, further performance improvement is limited due unwanted boundary effects of the circular structure. In this work, through an in‐depth study features’ characteristics, authors propose a novel strategy achieve simultaneous filter matching and regularization CNN when on fly. With...
In this paper, an object tracking approach is introduced for color video sequences. The presents the integration of distributions and probabilistic principal component analysis (PPCA) into particle filtering framework. Color are robust to partial occlusion, rotation scale invariant calculated efficiently. Principal Component Analysis (PCA) used update eigenbasis mean, which can reflect appearance changes tracked object. And a low dimensional subspace representation PPCA efficiently adapts...
Using remote sensing video to monitor aircraft dynamics is significant for military applications, airport management, and rescue. The has a fixed size obvious characteristics, so it suitable correlation filtering. Correlation filtering algorithms can extract features from input data predict motion trajectories, the calculation speed of filterings fast. Hence, such are advantageous tracking targets in images. In this article, an antidrift multifilter tracker based on filter Kalman proposed...