This repo. is an official implementation of the SCANet , which has been accepted in the journal Neural computing and applications, 2022.
An typical example
The main pipeline is shown as the following,
And qualitative and quantitative results are listed
pretrained ViT is available at (Baidu) [code:nnty]
python train.py
python test.py
We provide the trained model file (Baidu) [code:u4yy]
The saliency maps are also available (Baidu). [code:e0pi]
Please cite the SCANet in your publications if it helps your research:
@article{CHEN2022,
title = {Spatiotemporal Context-aware Network for Video Salient Object Detection},
author = {Tianyou Chen and Jin Xiao and Xiaoguang Hu and Guofeng Zhang and Shaojie Wang},
journal = {Neural Computing and Applications},
year = {2022},
}