ATSA

This codebase implements the system described in the paper:

Asymmetric Two-Stream Architecture for Accurate RGB-D Saliency Detection

Miao Zhang, Sun Xiao Fei, Jie Liu, Shuang Xu, Yongri Piao, Huchuan Lu. In ECCV 2020.

Prerequisites

• Ubuntu 18
• PyTorch 1.3.1
• CUDA 10.1
• Cudnn 7.5.1
• Python 3.7
• Numpy 1.17.3

Training and Testing Datasets

Training dataset

Download Link. Code: nx8x

Testing dataset

Download Link. Code: qqsf

Train/Test

test

Firstly, you need to download the 'Testing dataset' and the pretraind checpoint we provided (Baidu Pan. Code: d2o0). Then, you need to set dataset path and checkpoint name correctly. and set the param '--phase' as "test" and '--param' as 'True' in demo.py.

python demo.py

train

Once the train-augment dataset are prepared,you need to set dataset path and checkpoint name correctly. and set the param '--phase' as "train" and '--param' as 'True'(loading checkpoint) or 'False'(do not load checkpoint) in demo.py.

python demo.py

Contact Us

If you have any questions, please contact us (xiaofeisun@mail.dlut.edu.cn; 1605721375@mail.dlut.edu.cn).