Image to Sphere: Learning Equivariant Features for Efficient Pose Prediction.
David M. KleeOndrej BizaRobert PlattRobin WaltersPublished in: CoRR (2023)
Keyphrases
- image features
- extracting features
- low level
- original images
- image data
- prior knowledge
- single image
- input image
- image classification
- image content
- multiscale
- feature values
- keypoints
- feature set
- prediction accuracy
- linear predictors
- extracted features
- learning algorithm
- position and orientation
- visual object recognition
- image matching
- edge detection
- feature space
- feature extraction
- test images
- rotation invariant
- invariant features
- object detectors
- d objects
- global features
- image retrieval
- feature extractor
- image segmentation