Coarse-to-Fine Q-attention: Efficient Learning for Visual Robotic Manipulation via Discretisation.
Stephen JamesKentaro WadaTristan LaidlowAndrew J. DavisonPublished in: CVPR (2022)
Keyphrases
- coarse to fine
- efficient learning
- multiscale
- selective attention
- multiresolution
- object detection
- hierarchical segmentation
- manipulation tasks
- image registration
- scan line
- visual features
- dynamic programming
- active shape model
- membership queries
- learning algorithm
- matching scheme
- hierarchical representation
- multi view face detection
- visual attention
- image sequences
- image segmentation