An Hybrid CPU-GPU Parallel Multi-tracking Framework for Long-Term Video Sequences.
Juan P. D'AmatoLeonardo DominguezFranco StramanaAldo RubialesAlejandro PerezPublished in: WEA (2021)
Keyphrases
- graphics processing units
- tracking framework
- video sequences
- visual tracking
- face tracking
- object tracking
- parallel implementation
- gpu implementation
- parallel processing
- parallel computing
- graphics processors
- particle filtering
- appearance model
- multi target tracking
- particle filter
- motion estimation
- moving objects
- kalman filter
- memory bandwidth
- robust tracking
- dynamical model
- image sequences
- surveillance system
- video data
- multi view
- facial expressions
- video surveillance
- mean shift
- face detection
- level parallelism
- partial occlusion
- shared memory
- data association
- human motion
- background subtraction
- video frames
- viewpoint
- machine learning