Coarse-to-Fine Video Instance Segmentation With Factorized Conditional Appearance Flows.
Zheyun QinXiankai LuXiushan NieDongfang LiuYilong YinWenguan WangPublished in: IEEE CAA J. Autom. Sinica (2023)
Keyphrases
- coarse to fine
- active shape
- hierarchical segmentation
- multiscale
- deformable surface model
- hierarchical representation
- optical flow estimation
- appearance features
- active shape model
- deformable contour
- image pyramid
- multiresolution
- layered representation
- image registration
- object detection
- level set
- image segmentation
- video data
- dynamic textures
- bounding box
- segmentation algorithm
- unsupervised image segmentation
- dynamic programming
- image analysis
- shape prior
- segmentation method
- super resolution
- video sequences
- reinforcement learning
- object recognition
- video frames
- textured images
- energy function