https://github.com/bloc97/misr

Exploratory Project on Multiple-Image Super-Resolution
https://github.com/bloc97/misr

convolutional-neural-networks dlss super-resolution

Last synced: about 2 months ago
JSON representation

Exploratory Project on Multiple-Image Super-Resolution

• Host: GitHub
• URL: https://github.com/bloc97/misr
• Owner: bloc97
• License: mit
• Created: 2020-12-23T21:10:34.000Z (about 4 years ago)
• Default Branch: main
• Last Pushed: 2020-12-26T06:17:17.000Z (about 4 years ago)
• Last Synced: 2024-10-31T04:50:35.520Z (2 months ago)
• Topics: convolutional-neural-networks, dlss, super-resolution
• Language: Jupyter Notebook
• Homepage:
• Size: 4.36 MB
• Stars: 14
• Watchers: 4
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# MISR

Exploratory Project on Multiple-Image Super-Resolution for the Autumn 2020 IFT6145 Computer Vision course at Université de Montréal.

Explores the use of optical flow in mutiple-image super-resolution. This method should be similar to what is used by other projects such as the commonly known Nvidia DLSS 2.0.

## Resources

[Report](Report.pdf)  

[Presentation](Slides.pdf)  

Code (Jupyter Notebooks) located in the `Code` folder. (Uncleaned, will give you a headache...)  

The dataset used in this work is the [MPI Sintel Flow Dataset](http://sintel.is.tue.mpg.de/).  

## Results

![Examples](img/Comparison.png?raw=true)