AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

"Close, But Not There: Boosting Geographic Distance Sensitivity in Visual Place Recognition"

Sergio Izquierdo, Javier Civera

ECCV 2024 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 21 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train with CliqueMining batches on DINOv2 SALADDINOv2 SALAD · NordlandRecall@1: 76.0 -> 90.7Improved
train with CliqueMining batches on DINOv2 SALADDINOv2 SALAD · MSLS ChallengeRecall@1: 75.0 -> 82.7Improved
train with CliqueMining on MixVPRMixVPR · MSLS ChallengeRecall@1: 64.0 -> 65.6Improved
train with CliqueMining on MixVPRMixVPR · Pitts250k-testRecall@1: 94.6 -> 91.8Hurt
train with CliqueMining on DINOv2 SALADDINOv2 SALAD · Pitts250k-testRecall@1: 95.1 -> 95.2No clear effect
use CliqueMining rather than naively adding MSLS training data (ablation)DINOv2 SALAD · NordlandRecall@1: 68.27 -> 88.22Improved
use CliqueMining rather than naive GSV-Cities + MSLS (ablation)DINOv2 SALAD · MSLS Train (val subset)Recall@1: 79.96 -> 81.98Improved
keep online multi-similarity mining together with CliqueMining (ablation)CliqueMining · MSLS Train (val subset)Recall@1: 76.87 -> 81.98Improved
weighted random vs most-similar vs uniform random sequence sampling (ablation)CliqueMining graph creation · MSLS Train (val subset)Recall@1: 81.61 / 81.98 / 80.40No clear effect
recompute cliques every epoch (ablation)CliqueMining · MSLS Train (val subset)Recall@1: 81.98 -> 81.96No clear effect
set distance threshold tau to 25 instead of other values in 10-30 (ablation)CliqueMining · MSLS Train (val subset)Recall@1Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy