Hongqing Liu,
Here you can run EC-Fusion-RGBIR yourself using terminal command.
python -W ignore run.py config/RGB_image.yaml
Here you can change configuration file to load different datasets and using different network
dataset: 'RGB_image_single' # choose the datasets you want to use.
# Option: "single_RGB_image", "single_fusion_image", "multi_IR_image", ...
mode: 'visible' # 'IR', 'Fusion'
...
If you want to use network, you can add it in src/models.
Here I show how we store input dataset for single object detection, you can create a new file under file "datasets".
DATAROOT
└── datasets
└── file_name
├── thermal
│ ├── 00000.jpg
│ ├── 00001.jpg
│ ├── 00002.jpg
│ ├── ...
│ └── ...
├── visible
│ ├── 00000.jpg
│ ├── 00001.jpg
│ ├── 00002.jpg
│ ├── ...
│ └── ...
└── label.csv
Single object training framework
└── datasets
DATAROOT
└── datasets
└── file_name
├── thermal
│ ├── 00000.jpg
│ ├── 00001.jpg
│ ├── 00002.jpg
│ ├── ...
│ └── ...
├── visible
│ ├── 00000.jpg
│ ├── 00001.jpg
│ ├── 00002.jpg
│ ├── ...
│ └── ...
└── label.csv