Adversarial Reinforcement Learning With Object-Scene Relational Graph for Video Captioning.
Xia HuaXinqing WangTing RuiFaming ShaoDong WangPublished in: IEEE Trans. Image Process. (2022)
Keyphrases
- relational graph
- reinforcement learning
- video sequences
- video scene
- object detection and tracking
- object motion
- moving objects
- multiple objects
- stationary camera
- complex scenes
- moving camera
- object tracking
- dynamic scenes
- image frames
- d scene
- object models
- combining information from multiple
- multi label
- video surveillance
- video frames
- video data
- video images
- multi agent
- spatial relations
- image sequences
- target object
- video analysis
- space time
- visual input
- camera movement
- object segmentation
- dynamic textures
- d objects
- object detectors
- ground plane
- computer vision
- optical flow
- feature extraction
- decision trees