AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Region-Aware Pretraining for Open-Vocabulary Object Detection With Vision Transformers

Dahun Kim, Anelia Angelova, Weicheng Kuo

CVPR 2023 · Detection and segmentation › Open-set and few-shot detection · approach: Open-vocabulary object detection · 76 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
Use CPE with focal contrastive loss instead of standard full-image PE and softmax cross-entropy (ablation)RO-ViT pretraining · LVIS novel categoriesAP_r: 21.4 -> 24.3; +2.9 AP_rImproved
Add cropped positional embeddings (ablation)RO-ViT pretraining · LVIS novel categoriesAP_r: +2.4 AP_rImproved
Replace softmax cross-entropy with focal contrastive loss (ablation)RO-ViT pretraining · LVIS novel categoriesAP_r: +0.5 AP_rImproved
Remove positional embeddings (ablation)RO-ViT pretraining · LVIS novel categoriesAP_r: 21.4 -> 18.7Hurt
Replace learnable PE with sinusoidal PE (ablation)RO-ViT pretraining · LVIS novel categoriesAP_r: 21.4 -> 21.7No clear effect
Crop and resize the last feature map rather than positional embeddings (ablation)RO-ViT pretraining · LVIS novel categoriesAP_r: 21.4 -> 20.6Hurt
Add CPE and focal loss (ablation)Frozen RO-ViT backbone · LVIS novel categoriesAP_r: 9.7 -> 16.5; +6.8 AP_rImproved
Set backbone finetuning LR ratio to 0.1 (ablation)RO-ViT detector · LVIS novel categoriesAP_r: 16.5 -> 24.3Improved
Use localization-quality objectness in detector scoring (ablation)RO-ViT detector · LVIS novel categoriesAP_r: +2.0 AP_rImproved
Normalize classifier and mask output layers (ablation)RO-ViT detector · LVIS novel categoriesAP_r: +0.6 AP_rImproved
Use CPE and focal loss instead of baseline pretraining (ablation)RO-ViT ViT-B/16 · MS COCO retrievalimage-to-text Recall@1: 59.1 -> 62.2Improved
Increase backbone size from ViT-B/16 to ViT-L/16 (ablation)RO-ViT with focal and CPE · MS COCO retrievalimage-to-text Recall@1: 62.2 -> 67.0Improved

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