Construction worker hardhat-wearing detection based on an improved BiFPN
The code is Construction worker hardhat-wearing detection based on an improved BiFPN paper's test code. To run the code, you need to do the following steps:
- Create checkpoint folder in the root path.
- Download the trained model from https://pan.baidu.com/s/15mBq-_snJBnhL6LY_EP-sA (password: sff2) and put it in the checkpoint folder.
- Create the dataset folder under the ./data/.
- Write the corresponding data of the test dataset in ./dataset/. Data format: image_ path x_ min,y_ min,x_ max,y_ max,class_ id ...
- Run evaluate.py
- Run mAP.py. You will get result of the mAP.