Enhance Low Resolution Image using GANs
Mentors :
Hitaansh Shah
Vidit Khazanchi
Mentees :
7-8
The project involves implementation of ESRGAN using PyTorch to enhance low resolution images having a lot of loss in details to images having great spatial resolution. If time permits , we shall build a basic web app to deploy the project . It will be extremely fun to learn about GANs and you shall gain a lot of insight on deep learning in general. One can check out https://youtu.be/WCAF3PNEc_c . Our project shall be on the same lines. You may also go through the Andrew NG course on CNN https://youtube.com/playlist?list=PLkDaE6sCZn6Gl29AoE31iwdVwSG-KnDzF. The research paper that we will largely be following https://arxiv.org/abs/1809.00219 .
Prerequisites:
Intermediate Python Skills and a lot of Enthusiasm to learn about Deep Learning and Neural Networks.
Tentative Timeline :
| Week | Work |
|---|---|
| Week 1 | Brush up Python Programming and OOPs (Will be useful while building models) |
| Week 2 | Learn about Neural Networks and a bit of CNNs as well. |
| Week 3 | More CNN and learning PyTorch. |
| Week 4 | Implementing Neural Networks using Pytorch and learning about GANs. |
| Week 5 | Reading the paper on ESRGAN. |
| Week 6 | Start Implementing the model on the lines of the architecture as mentioned in the paper. |
| Week 7 | Finish Coding the same and debugging. |
| Week 8 | Build a Basic Website to deploy the project . |
