AI Paper Atlas

Research map / Detection and segmentation / CNN features

Cheaper Pre-training Lunch: An Efficient Paradigm for Object Detection

Dongzhan Zhou, Xinchi Zhou, Hongwen Zhang, Shuai Yi, Wanli Ouyang

ECCV 2020 · Detection and segmentation › CNN features · approach: Region-based and query-based detectors · 20 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
replace ImageNet pre-training with Montage pre-training at 1/4 costFaster R-CNN C4 ResNet-50 · COCO val2017AP: 34.8 -> 36.3Improved
replace ImageNet pre-training with Montage pre-trainingFaster R-CNN FPN ResNet-50 · COCO val2017AP: 36.2 -> 36.5No clear effect
replace ImageNet pre-training with Montage pre-trainingMask R-CNN FPN ResNet-50 · COCO val2017AP: 37.3 -> 37.5No clear effect
replace ImageNet pre-training with Montage pre-trainingFaster R-CNN ResNet-101 · COCO val2017AP: 38.3 -> 39.2Improved
replace ImageNet pre-training with Montage pre-trainingFaster R-CNN ResNeXt101-32x4d · COCO val2017AP: 40.2 -> 40.2No clear effect
use Montage pre-training instead of training detector from scratch (GN), 1x scheduleFaster R-CNN ResNet-50 · COCO val2017AP: 28.6 -> 36.3Improved
use Montage pre-training instead of training detector from scratch (GN), 2x scheduleFaster R-CNN ResNet-50 · COCO val2017AP: 32.6 -> 37.5Improved
add Montage assembly and ERF-adaptive loss to MS-COCO classification pre-training (1.73 GPU days) (ablation)MS-COCO pre-training · COCO val2017AP: 33.5 -> 36.3Improved
extend MS-COCO classification pre-training to ImageNet cost without Montage (ablation)MS-COCO w/o Montage · COCO val2017AP: 33.5 -> 34.4No clear effect
use pad and crop rather than warp to fit samples (ablation)Montage pre-training · COCO val2017AP: 34.6 -> 35.2Improved
use pad and crop rather than resize to fit samples (ablation)Montage pre-training · COCO val2017AP: 34.7 -> 35.2Improved
use ERF-adaptive dense classification instead of global classification (ablation)Montage pre-training · COCO val2017AP: 34.3 -> 36.3Improved

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