VITA: Video Instance Segmentation via Object Token Association.
Miran HeoSukjun HwangSeoung Wug OhJoon-Young LeeSeon Joo KimPublished in: CoRR (2022)
Keyphrases
- object segmentation
- video objects
- video segmentation
- multiple objects
- video sequences
- video scene
- video data
- motion cues
- bounding box
- segmentation algorithm
- image segmentation
- object motion
- video streams
- signed distance
- video analysis
- d objects
- moving objects
- object segmentation and tracking
- video content
- level set
- segmentation method
- background subtraction
- foreground background segmentation
- background pixels
- object shape
- video clips
- object tracking
- image regions
- multimedia
- pixel level
- region growing
- video surveillance
- object model
- key frames
- point trajectories
- final segmentation
- video tracking
- image segments
- object contours
- edge detection
- shape prior
- medical images
- video frames
- global shape
- object detection
- video shots
- spatial and temporal
- space time
- multiscale
- objects in video sequences
- target object