Cooperative-Students: Navigating Unsupervised Domain Adaptation in Nighttime Object Detection

Official implementation of Cooperative-Students: Navigating Unsupervised Domain Adaptation in Nighttime Object Detection.

Installation

Prerequisites

• Python ≥ 3.6

• PyTorch ≥ 1.5 and torchvision that matches the PyTorch installation.

• Detectron2 == 0.6 (The version I used to run my code)

Our tested environment

• 2 Nvidia A40 (2 * 9 batch size)

Install python env

To install required dependencies on the virtual environment of the python (e.g., virtualenv for python3), please run the following command at the root of this code:

$ python3 -m venv /path/to/new/virtual/environment/.
$ source /path/to/new/virtual/environment/bin/activate

For example:

$ mkdir python_env
$ python3 -m venv python_env/
$ source python_env/bin/activate

Build Detectron2 from Source

Follow the INSTALL.md to install Detectron2.

Dataset download

1. Download the datasets

   Dataset	URL	Num of Classes	Daytime	Nighttime (UDA)	Nighttime (Eval)
   BDD100K	Link	10	36,728	32,998	4,707
   SHIFT	Link	6	19,452	8,497	1,200
   ACDC	Link	8	1,200	400	400

2. Organize the dataset in COCO format following:

cos/
└── data/
    └── bdd100k/
        └── images/
            ├── train/
            └── val/
        ├── train_day.json
        └── train_night.json
        └── val_night.json
   └── shift/
        └── images/
            ├── train/
            └── val/
        ├── train_day.json
        └── train_night.json
        └── val_night.json
   └── ACDC/
        ├── rgb_anon/
        └── instancesonly_day_train_gt_detection_continue.json
        └── instancesonly_night_train_gt_detection_continue.json
        └── instancesonly_night_val_gt_detection_continue.json

Training

• Train the Cooperative Students from daytime to nighttime using BDD100K

python train_net.py \
      --num-gpus 2 \
      --config configs/faster_rcnn_R50_BDD.yaml \
      OUTPUT_DIR output/bdd100k

• Train the Cooperative Students from daytime to nighttime using SHIFT

python train_net.py \
      --num-gpus 2 \
      --config configs/faster_rcnn_R50_SHIFT.yaml \
      OUTPUT_DIR output/shift

Resume the training

python train_net.py \
      --resume \
      --num-gpus 2 \
      --config configs/faster_rcnn_R50_BDD.yaml MODEL.WEIGHTS <your weight>.pth

Evaluation

python train_net.py \
      --eval-only \
      --num-gpus 2 \
      --config configs/faster_rcnn_R50_BDD.yaml \
      MODEL.WEIGHTS <your weight>.pth

Citation

@article{yuan2024cooperative,
  title={Cooperative Students: Navigating Unsupervised Domain Adaptation in Nighttime Object Detection},
  author={Yuan, Jicheng and Le-Tuan, Anh and Hauswirth, Manfred and Le-Phuoc, Danh},
  journal={arXiv preprint arXiv:2404.01988},
  year={2024}
}