BSUV-Net 2.0: Spatio-Temporal Data Augmentations for Video-Agnostic Supervised Background Subtraction.
Mustafa Ozan TezcanPrakash IshwarJanusz KonradPublished in: IEEE Access (2021)
Keyphrases
- background subtraction
- spatio temporal data
- moving objects
- video surveillance
- video sequences
- moving camera
- spatio temporal
- dynamic scenes
- foreground background segmentation
- dynamic background
- image sequences
- foreground objects
- motion detection
- background model
- moving object detection
- video frames
- background modeling
- video streams
- object detection
- gaussian mixture model
- space time
- video data
- appearance model
- sensor data
- background image
- spatial databases
- object tracking
- camera motion
- infrared
- spatial data
- computer vision
- post processing
- semi supervised
- feature selection
- data streams
- spatial features
- machine learning
- temporal information
- pairwise
- nearest neighbor
- motion patterns
- data sets
- human actions
- event detection