A deep learning web application to boost the resolution of images.
Please use this link to view a short set of pdf slides that walks through the functionality and development process of PixelBoost.
The project is hosted using Heroku and GitHub Pages at https://pixelboost.github.io
This repository showcases the three main parts I had to develop for the application: the Jupyter notebook where I developed the model, a folder containing the Flask backend, a folder containing the Flutter frontend, and a folder containing the Docker image and model files responsible for model computations.
The interactive python notebook where I utilized TensorFlow and numpy to architect, train, test, and compile the convolutional neural network that performs image super resolution.
The project for the Flask app responsible for image pre/post-processing. See srcnnimg/app.py and srcnnimg/model.py for the Python code I wrote.
The Flutter project used for the website. See pixelboost/lib/main.dart and pixelboost/lib/widgets.dart for the main Dart code I wrote.
The files used for a TF Serving Docker image for model predictions.