AsymLoc uses teacher-student distillation with geometry-driven matching to enable efficient nearest-neighbor feature matching, achieving 95% of teacher accuracy with 10x smaller models on localization benchmarks.
Compatibility- aware heterogeneous visual search
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
AsymLoc: Towards Asymmetric Feature Matching for Efficient Visual Localization
AsymLoc uses teacher-student distillation with geometry-driven matching to enable efficient nearest-neighbor feature matching, achieving 95% of teacher accuracy with 10x smaller models on localization benchmarks.