shnew / R2IPoints

Codes for R2IPoints : Pursuing Rotation-Insensitive Point Representation for Aerial Object Detection.

R2IPoints : Pursuing Rotation-Insensitive Point Representation for Aerial Object Detection

Codes for R2IPoints : Pursuing Rotation-Insensitive Point Representation for Aerial Object Detection.

The overview of R2IPoints:

Getting Started

Prerequisties

• pip install torch==1.1.0
• pip install torchvision==0.3.0
• pip install mmdet==1.2.0
• pip install mmcv==0.3.1
• pip install -r requirements/build.txt
• pip install -v -e .

Datasets preparation

• DIOR: http://www.escience.cn/people/gongcheng/DIOR.html
• DOTA: https://captain-whu.github.io/DOTA/
  • code for cutting images: https://github.com/dingjiansw101/AerialDetection/tree/master/DOTA_devkit

Quick start

• python tools/train.py ./configs/R2IPoints/R2IPoints_moment_r101_fpn_2x_DIOR.py
• python tools/test.py ./configs/R2IPoints/R2IPoints_moment_r101_fpn_2x_DIOR.py ${CHECKPOINT_FILE} --eval mAP

Pre-trained models

• https://download.pytorch.org/models/resnet101-5d3b4d8f.pth