1. Requirements

Python 3.7+, Pytorch 1.9.0, Cuda 11.1, TensorboardX 2.4, opencv-python

For detailed environment configuration, please refer to "./environment/env.txt" (for pip) and "./environment/env.yaml" (for conda).

2. Training & Testing

• Train the model:

  bash train_best_model.sh

• Test the model:

  bash test_model.sh

  The predicted saliency maps will be saved in "./Our_Proposed_Best_Model/pred_maps". The metrics results will be saved in "./Our_Proposed_Best_Model/score/result.txt"

3. single-bounding-box annotations:

Our proposed annotations for the trainset can be downloaded here.