SeqLoc aggregates per-frame pose likelihoods from OSM-based geo-localization into an online log-belief, using entropy weighting, map-guided recovery, and sub-grid smoothing; on the new CV-FSS benchmark it raises RHO recall at 5 m from 10.6% to 75.0%.
Computer Science Review , volume=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SeqLoc: Beyond the Single Frame for Cross-View Geo-Localization in Feature-Sparse Scenes
SeqLoc aggregates per-frame pose likelihoods from OSM-based geo-localization into an online log-belief, using entropy weighting, map-guided recovery, and sub-grid smoothing; on the new CV-FSS benchmark it raises RHO recall at 5 m from 10.6% to 75.0%.