Aligned to the Object, Not to the Image: A Unified Pose-Aligned Representation for Fine-Grained Recognition.
Pei GuoRyan FarrellPublished in: WACV (2019)
Keyphrases
- fine grained
- object representation
- invariant recognition
- partial occlusion
- d objects
- object models
- object recognition and pose estimation
- image representation
- coarse grained
- spatial relations
- improve recognition accuracy
- object recognition
- position and orientation
- recognizing objects
- object shape
- recognition and pose estimation
- object representations
- pose estimation
- invariant representation
- multiple objects
- image features
- image classification
- object tracking
- pose determination
- image regions
- input image
- three dimensional objects
- access control
- lighting conditions
- tightly coupled
- object appearance
- keypoints
- objects in cluttered scenes
- pose space
- rigid objects
- image retrieval
- bounding box
- data lineage
- image matching
- object pose
- single image
- moving objects
- face representation
- feature points
- target object
- appearance variations
- object class
- illumination variations