/Saliency-Prediction-using-GAN

This is a SalGAN implementation without the pre-trained VGG features written in PyTorch.

Primary LanguageJupyter Notebook

Saliency Prediction using GAN

This is a SalGAN implementation without the pre-trained VGG features written in PyTorch.

Report

For all other information and a detailed report and metrics, see Final Report.pdf

Usage

Clone the repo. Download the trained generator weights here. Run Test_Model.ipnyb using Jupyter Notebook and use functions saliency_from_link and saliency_from_val

Train it yourself

Step 1:

Download the images and fixation maps from SALICON and save this data in a folder called data.

Step 2:

Clone this repository. Install dependencies.

pip install -r requirements.txt

Step 3:

Rezise the images into 256x192

python resize_data.py

Step 4:

Train the model

python main.py