Long and Short Memory Balancing in Visual Co-Tracking using Q-Learning.
Kourosh MeshgiMaryam Sadat MirzaeiShigeyuki ObaPublished in: CoRR (2019)
Keyphrases
- cooperative
- particle filter
- reinforcement learning
- long sequences
- real time
- visual information
- particle filtering
- function approximation
- kalman filter
- appearance model
- multi agent
- visual features
- visual tracking
- narrow field of view
- memory space
- object tracking
- state space
- low level
- high level
- memory requirements
- learning rate
- memory usage
- working memory
- memory size
- video surveillance
- action selection
- dynamic programming
- short term memory
- multi agent reinforcement learning