A5068126521- Neonatal and fetal brain pathology
- Human-Automation Interaction and Safety
- Advanced Vision and Imaging
- Gaze Tracking and Assistive Technology
- Sleep and Work-Related Fatigue
- Human Pose and Action Recognition
- Anesthesia and Neurotoxicity Research
- Video Surveillance and Tracking Methods
- Traumatic Brain Injury and Neurovascular Disturbances
China University of Mining and Technology
2020
Ministry of Education of the People's Republic of China
2020
Soochow University
2012
In view of the fact that detection driver’s distraction is a burning issue, this study chooses head pose as evaluation parameter for driving and proposes driver method based on pose. The effects single regression classification combined with are compared in terms accuracy, four kinds classical networks improved trained using 300W-LP AFLW datasets. HPE_Resnet50 best accuracy selected estimator applied to ten-category distracted dataset SF3D obtain 20,000 sets data. differences between classes...
Although correlation filter-based trackers have demonstrated excellent performance for visual object tracking, there remain several challenges to be addressed. In this work, we propose a novel tracker based on the filter framework. Traditional face difficulty in accurately adapting changes scale of target when moves quickly. To address this, suggest adaptive scheme prediction scales. We also incorporate speed-based model update method further improve overall tracking performance. Experiments...