AI Paper Atlas

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

CPlaNet: Enhancing Image Geolocalization by Combinatorial Partitioning of Maps

Paul Hongsuck Seo, Tobias Weyand, Jack Sim, Bohyung Han

ECCV 2018 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 78 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
combine classifiers via combinatorial partitioning (CPlaNet[1-5] vs naive averaging)ClassSet 1-5 ensemble · Im2GPS3kaccuracy within 1 km (%): 8.5 -> 9.9Improved
combinatorial partitioning of two classifiers (CPlaNet[1-2] vs Average[1-2])ClassSet 1-2 ensemble · Im2GPS3kaccuracy within 1 km (%): 8.2 -> 9.3Improved
use five classifiers instead of two in combinatorial partitioning (ablation)CPlaNet · Im2GPS3kaccuracy within 1 km (%): 9.3 -> 9.9Improved
add reproduced PlaNet as an additional base classifier (ablation)CPlaNet[1-5] · Im2GPS3kaccuracy within 1 km (%): 9.9 -> 10.2Improved
add reproduced PlaNet as an additional base classifier (ablation)CPlaNet[1-5] · Im2GPS3kaccuracy within 2500 km (%): 64.6 -> 64.6No clear effect
use visual-distance edge weights instead of geographic-only edge weights for geoclass generation (ablation)ClassSet · Im2GPS3kaccuracy within 1 km (%): 8.0 -> 8.4Improved
use PlaNet-style ImageNet features instead of geolocation-trained features (ImageNetFeat vs ClassSet 1) (ablation)ClassSet 1 · Im2GPS3kaccuracy within 1 km (%): 8.4 -> 3.0Hurt
CPlaNet[1-5,PlaNet] versus PlaNet (reprod) at coarse scaleCPlaNet · YFCC4kaccuracy within 1 km (%): 5.6 -> 7.9Improved
CPlaNet[1-5,PlaNet] versus PlaNet (reprod) at 25 kmCPlaNet · YFCC4kaccuracy within 25 km (%): 14.3 -> 14.8No clear effect
CPlaNet[1-5,PlaNet] versus PlaNet (reprod) at 2500 kmCPlaNet · Im2GPSaccuracy within 2500 km (%): 81.9 -> 78.5Hurt
CPlaNet[1-5,PlaNet] versus PlaNet (reprod) at street levelCPlaNet · Im2GPSaccuracy within 1 km (%): 11.0 -> 16.5Improved
add five classification heads instead of oneCPlaNet · Im2GPS3kinference time per image (s): 0.12 -> 0.13No clear effect

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