An Approach to Overcome Occlusions in Visual Tracking: By Occlusion Estimating Agency and Self-Adapting Learning Rate for Filter's Training.
Kaiwen JiangFeng QianCe SongBao ZhangPublished in: IEEE Signal Process. Lett. (2018)
Keyphrases
- visual tracking
- learning rate
- partial occlusion
- multilayer neural networks
- tracking objects
- training algorithm
- appearance variations
- adaptive learning rate
- robust tracking
- background clutter
- particle filter
- particle filtering
- appearance model
- object appearance
- convergence rate
- mean shift
- learning algorithm
- unscented kalman filter
- object tracking
- video sequences
- object class
- rapid convergence
- neural network
- data association
- illumination variations
- convergence speed
- articulated structures
- interacting multiple model
- computer vision
- pose variations
- markov random field
- multi class
- support vector machine
- multi objective
- image sequences
- genetic algorithm