Overview / Usage

Deep Learning Model that can restore image, upto 4x improvement in resolution quality.

Methodology / Approach

We are using a VGG as the base model and then add a custom head over the top, taking features from Enhanced Deep Residual Networks for Super Resolution (EDSR). We used Pixel shuffling and ICNR initialization

Technologies Used

DevCloud, Jupyter Notebook

Repository

https://github.com/prajjwal1/super_resolution