ysyscool / DINet

A dilated inception network for visual saliency prediction (TMM 2019)

DINet

This repository contains the reference code for our TMM paper: arXiv Paper Version

If you use any part of our code, or DINet is useful for your research, please consider citing::

@article{yang2019dilated,
  title={A dilated inception network for visual saliency prediction},
  author={Yang, Sheng and Lin, Guosheng and Jiang, Qiuping and Lin, Weisi},
  journal={IEEE Transactions on Multimedia},
  volume={22},
  number={8},
  pages={2163--2176},
  year={2019},
  publisher={IEEE}
}

Requirements

• Python 2.7
• Keras 2.1.2
• Tensorflow-gpu 1.3.0
• opencv-python

Getting Started

Installation

• Clone this repo:

git clone https://github.com/ysyscool/DINet
cd DINet
mkdir models

• Download weights from Google Drive. Put the weights into

cd models

Train/Test

Download the SALICON 2015 dataset and modify the paths in config.yaml And then using the following command to train the model

python main.py --phase=train --batch_size=10

For testing, modify the variables of weightfile (in line 217) and imgs_test_path (in line 220) in the main.py. And then using the following command to test the model

python main.py --phase=test

Evaluation on SALICON dataset

Please refer to this link.

Acknowledgments

Code largely benefits from sam.