ec-classifer-training-framework's Introduction

EC-classifier-training-framework

Hongqing Liu,

Run

Here you can run EC-Fusion-RGBIR yourself using terminal command.

python -W ignore run.py config/RGB_image.yaml

configuration file

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'
...

network

If you want to use network, you can add it in src/models.

Datasets

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

Recommend Projects