Skip to content
/ SPE Public

Official implementation of ECCV2022 paper End-to-End Weakly Supervised Object Detection with Sparse Proposal Evolution

Notifications You must be signed in to change notification settings

MingXiangL/SPE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

End-to-End Weakly Supervised Object Detection with Sparse Proposal Evolution

Install

pip install torch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0
pip install timm
pip install opencv-python
pip install scipy
pip install pycocotools

Data Preperation

We support COCO format. For VOC format annotations, please transfer them to coco format first.

For example, in experiments on VOC0712 trainval first transfer annotations of VOC2007 trainval and VOC2012 trainval to coco format in a single json file. # note: to evaluate mAP performance, please use the original annotation of voc07 test.

Models

VOC0712: mAP=51.0

COCO2017: AP=7.9, AP50=19.5, AP75=5.5

Train & Val

COCO 2017:

python scripts/run_coco17.py

VOC 07+12

python scripts/run_voc0712.py

About

Official implementation of ECCV2022 paper End-to-End Weakly Supervised Object Detection with Sparse Proposal Evolution

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages