AI Paper Atlas

Research map / Detection and segmentation / CNN features

Online Data Curation for Object Detection via Marginal Contributions to Dataset-level Average Precision

Zitang Sun, Masakazu Yoshimura, Junji Otsuka, Atsushi Irie, Takeshi Ohashi

CVPR 2026 · Detection and segmentation › CNN features · approach: Efficient DETR query design · 0 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
add DetGain online sampling with strong augmentationFaster R-CNN R50 · COCO val2017AP: 37.5 -> 40.0Improved
add DetGain online sampling with strong augmentationDeformable DETR R50 · COCO val2017AP: 46.6 -> 48.9Improved
strong data augmentation alone (ablation)ATSS R50 · COCO val2017AP: 39.2 -> 38.6Hurt
strong data augmentation alone (ablation)Faster R-CNN R50 · COCO val2017AP: 37.4 -> 37.5No clear effect
add strong augmentation to DetGain sampling (ablation)Faster R-CNN R50 · COCO val2017AP: 37.3 -> 39.9Improved
DetGain sampling without augmentation (ablation)Faster R-CNN R50 · COCO val2017AP: 37.4 -> 37.3No clear effect
use larger teacher backbone (Res50 to Res152) (ablation)DetGain Faster R-CNN · COCO val2017AP: 39.6 -> 40.0Improved
DetGain scoring instead of loss-learnability (ablation)Faster R-CNN R50 · COCO val2017AP: 38.9 -> 40.0Improved
DetGain scoring instead of loss-learnability (ablation)FCOS R50 · COCO val2017AP: 38.1 -> 40.9Improved
loss-based hard mining instead of uniform sampling (ablation)FCOS R50 · COCO val2017AP: 38.2 -> 34.5Hurt
combine DetGain with CrossKD (ablation)FCOS R50 (2x) · COCO val2017AP: 41.1 -> 42.2Improved
use uniform Beta(1,1) prior instead of fitted score distributions (ablation)DetGain · simulated super-batchesSpearman rho: mean 0.94No 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