Find Objects and Focus on Highlights: Mining Object Semantics for Video Highlight Detection via Graph Neural Networks.
Yingying ZhangJunyu GaoXiaoshan YangChang LiuYan LiChangsheng XuPublished in: AAAI (2020)
Keyphrases
- object detection and tracking
- neural network
- bounding box
- multiple objects
- multi object
- object detectors
- video scene
- object model
- d objects
- detecting objects
- object detection
- visual objects
- data objects
- cluttered scenes
- partial occlusion
- individual objects
- object identity
- moving objects
- focus of attention
- similar objects
- object tracking
- object segmentation
- complex scenes
- video sequences
- object motion
- cluttered background
- video analysis
- video objects
- object classes
- method invocation
- object orientation
- object level
- motion trajectories
- object features
- multimedia
- crowded scenes
- video data
- event detection
- background clutter
- target object
- dynamic scenes
- stationary camera
- background subtraction
- object models
- topological information
- itemsets
- complex background
- face detection and tracking
- keypoints
- mobile objects
- spatial relations
- spatial relationships
- digital objects
- graph data
- visual input
- video content
- complex objects
- relational structures
- graph patterns
- space time
- temporal continuity
- quasi cliques
- pattern mining
- object shapes
- video surveillance
- raw image