smART is a project developed for Hack the North 2021. Designed to be give art recommendations through the form of a web app that uses the InceptionV3 CNN Model to output visually similar images compared to a user inputted image. Developed using Python, Flask, TensorFlow, and AWS EC2, this was a challenging but rewarding project to develop in 48 hours.
The smART web application uses a Python backend and HTML/CSS frontend, and utilizes the Flask Python package to integrate them together. To give recommendations, I used the Inception V3 CNN and trained it on AWS EC2 using a dataset I found online at https://www.kaggle.com/ikarus777/best-artworks-of-all-time.
Some features I would like to develop and implement in the future are:
- Cloud storage to reduce load on machine
- Optimization of code to process results faster
- Redesign the layout to have a better user experience
This has been a fantastic learning experience for everyone involved and we all look forward to seeing where this project could lead to in the future!