jokestv2 / image_fusion_for_sr

Image fusion via an autoencoder with dense-conv-block. (Pytorch)

📝 Table of Contents

• About
• To Course TA
• Usage

About

This is the course project for CSC2503 at UofT. The title of the project is "Image Fusion for Multi-Reference-based Super-Resolution".

To CSC2503 Course TA

The best way to evaluation the amount of work I have contributed to this project is to look at the commit history of this GitHub repository. The commit message serves as a good summary of my code and the total lines of code modified are also clearly shown.

Usage

First, download the dataset from my Google Drive (link) and unpack it into a folder named 'data' in the root directory. This dataset is generated by a SRefSR named C2-matching (https://github.com/lastbasket/C2-Matching).

To test the model on the dataset, run:

python test --data_root="./data/CUFED5_fusion" --batch_size=1 

To fine-tune the model on the dataset, run:

python train --data_root="./data/CUFED5_fusion" --batch_size=8 --lr=1e-4