RIFR

This project provides an implementation for "Balancing Attention to Base and Novel Categories for Few-Shot Object Detection in Remote Sensing Imagery" on PyTorch.

Requirements

• Python 3.7+
• PyTorch 1.5+
• mmcv 1.3.12+
• mmdet 2.16.0+
• mmcls 0.15.0+

Get Started

step1: base training

bash ./tools/detection/dist_train.sh \
    configs/detection/rifr/dior/split1/rifr_r101_fpn_dior-split1_base-training.py 2

step2: reshape the bbox head of base model

python -m tools.detection.misc.initialize_bbox_head \
    --src1 work_dirs/rifr_r101_fpn_dior-split1_base-training/latest.pth \
    --method randinit \
    --save-dir work_dirs/rifr_r101_fpn_dior-split1_base-training

step3: few shot fine-tuning

bash ./tools/detection/dist_train.sh \
    configs/detection/rifr/dior/split1/rifr_r101_fpn_pt-loss_dior-split1_5shot-fine-tuning.py 2

Acknowledgement

This repo is based on the open-source mmfewshot project. We appreciate all the contributors who participated in the project.