End-to-end video background subtraction with 3d convolutional neural networks.
Dimitrios SakkosHeng LiuJungong HanLing ShaoPublished in: Multim. Tools Appl. (2018)
Keyphrases
- end to end
- background subtraction
- convolutional neural networks
- video surveillance
- moving camera
- dynamic scenes
- object detection and tracking
- foreground background segmentation
- dynamic background
- scalable video
- moving objects
- foreground objects
- object detection
- image sequences
- video sequences
- motion detection
- background model
- background modeling
- gaussian mixture model
- foreground detection
- video streams
- video frames
- post processing
- video data
- surveillance videos
- moving object detection
- background image
- appearance model
- video content
- gait recognition
- shadow detection
- detection of moving objects
- multimedia
- real time
- foreground regions
- congestion control
- video analysis
- space time
- pixel wise
- feature space