• Python3.8

• Pytorch1.8.2 LTS

• Cuda11.1

• OpenCV-Python

Download the following datasets and unzip them into dataset folder

• VT821
• VT1000
• VT5000

First, Install the required third-party libraries in therequirements.txt

pip install -r requirement.txt

Second, please download the pretrained model into run/models folder.

• model: https://pan.baidu.com/s/1Ss5w-vRNQAhiPX1er1mDfQ?pwd=2edr [password: 2edr]

Third, Test the model and get the predicted saliency maps, which will be saved into run/maps folder.

python test.py

Fixed the related configuration in the config.py file，We recommend using the pre-training model we provide for the best training accuracy

python train.py

• saliency maps: https://pan.baidu.com/s/1MRNDrx-J3gYFZDs_8bFy9g?pwd=suep [password: suep]
• Trained models: https://pan.baidu.com/s/1Ss5w-vRNQAhiPX1er1mDfQ?pwd=2edr [password: 2edr]