PETRI: Reducing Bandwidth Requirement in Smart Surveillance by Edge-Cloud Collaborative Adaptive Frame Clustering and Pipelined Bidirectional Tracking.
Ruoyang LiuLu ZhangJingyu WangHuazhong YangYongpan LiuPublished in: DAC (2021)
Keyphrases
- clustering method
- clustering algorithm
- k means
- real time
- cluster analysis
- visual tracking
- consecutive frames
- image frames
- cloud computing
- edge detection
- particle filtering
- moving object extraction
- temporal continuity
- single frame
- network bandwidth
- collaborative learning
- data flow
- edge information
- frame rate
- motion analysis
- hierarchical clustering
- particle filter
- unsupervised learning
- kalman filter
- motion segmentation
- document clustering
- moving target
- end to end
- motion model
- successive frames
- high dimensional data
- anomaly detection