A cascade of feature, visibility, and geometry filters lets image-based localization work from compact binary descriptors with competitive accuracy and much lower memory.
Toroidal con- straints for two-point localization under high outlier ratios
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Cascaded Parallel Filtering for Memory-Efficient Image-Based Localization
A cascade of feature, visibility, and geometry filters lets image-based localization work from compact binary descriptors with competitive accuracy and much lower memory.