GeoMix achieves new state-of-the-art results in descriptor-free 2D-3D matching by adding directional embeddings, learnable global context nodes, and multi-detector training, cutting rotation and translation errors by up to 90% on standard benchmarks.
arXiv preprint arXiv:2508.18971 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
A query-based attack reconstructs substantial 3D geometry and approximate appearance from scene coordinate regression models, contradicting prior privacy-preserving claims.
citing papers explorer
-
GeoMix: Descriptor-Free Visual Localization via Global Context and Multi-Detector Training
GeoMix achieves new state-of-the-art results in descriptor-free 2D-3D matching by adding directional embeddings, learnable global context nodes, and multi-detector training, cutting rotation and translation errors by up to 90% on standard benchmarks.
-
Seeing Through the Weights: Privacy Leakage in Scene Coordinate Regression
A query-based attack reconstructs substantial 3D geometry and approximate appearance from scene coordinate regression models, contradicting prior privacy-preserving claims.