ART is a neural network that iteratively estimates coarse-to-fine image transformations via an auto-regressive pipeline with hierarchical features and cross-attention guidance for robust alignment under sparse features and large deformations.
Surf: Speeded up robust features
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2representative citing papers
Learning rotation invariance in descriptors matches the performance of matcher-level invariance but allows earlier invariance, faster matchers, and no loss in upright performance when trained at scale.
citing papers explorer
-
Auto-regressive transformation for image alignment
ART is a neural network that iteratively estimates coarse-to-fine image transformations via an auto-regressive pipeline with hierarchical features and cross-attention guidance for robust alignment under sparse features and large deformations.
-
Who Handles Orientation? Investigating Invariance in Feature Matching
Learning rotation invariance in descriptors matches the performance of matcher-level invariance but allows earlier invariance, faster matchers, and no loss in upright performance when trained at scale.