AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

On Label Granularity and Object Localization

Elijah Cole, Kimberly Wilber, Grant Van Horn, Xuan Yang, Marco Fornoni, Pietro Perona, Serge Belongie, Andrew Howard, Oisin Mac Aodha

ECCV 2022 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised object localization · 16 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 on coarser label granularity instead of species labelsCAM · iNatLoc500MaxBoxAccV2: +5.1 MaxBoxAccV2Improved
train on coarser label granularity instead of species labelsCutMix · iNatLoc500MaxBoxAccV2: +6.6 MaxBoxAccV2Improved
coarsen labels all the way to phylumWSOL methods · iNatLoc500MaxBoxAccV2Hurt
train on coarser labels (manufacturer/family instead of variant)CAM · FGVC-AircraftMaxBoxAccV2: +4.3 MaxBoxAccV2Improved
train on coarser labelsCAM · CUBMaxBoxAccV2: +4.4 MaxBoxAccV2Improved
coarsen labels using ImageNet hierarchyCAM · ImageNetMaxBoxAccV2No clear effect
aggregate localization outputs of similar fine categories (CAM-Agg) instead of vanilla CAMCAM · iNatLoc500MaxBoxAccV2Hurt
train at family level with 50 images per categoryCAM · iNatLoc500MaxBoxAccV2: 9200 images matches full species-level CAM (~15x fewer images)Improved
30 hyperparameter sets instead of 5 (ablation)CAM species-level · iNatLoc500MaxBoxAccV2: 60.2 -> 60.8Improved
train 10 epochs instead of 2 (ablation)CAM · iNatLoc500MaxBoxAccV2Improved
train with FSL-Det few-shot detection on full-supervision boxes instead of WSOLFSL-Det vs CAM · iNatLoc500MaxBoxAccV2: 60.2 -> 83.6Improved

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