Inspiration
An interest in the recent commercialization of data collection.
What it does
Prompts the user to draw an image based off a given word, and evaluates the drawing based on the accuracy. The program then further takes this evaluation to rank them against users across the globe, or against friend users that use the same starter code as them.
How we built it
The website was built in HTML, CSS, and JavaScript. No framework such as React was used because we did not determine that a framework was needed. Firebase was used as the realtime database that the JavaScript could connect to for storing data (similarity scores of the users). We split up the work between all team members but stayed in constant communication to prevent any merge conflicts down the road.
Challenges we ran into
There were some general aesthetic and design issues when it came to the website, but the group was able to eventually decide on a retro theme. Moreover, the machine learning algorithms had to be fiddled with in order to assure the greatest accuracy. Connecting to the ranking base was also tedious and required a good bit of code in order to properly store results and give the user ranking feedback depending on the mode.
Accomplishments that we're proud of
We are proud that we were able to host the game and import it into another website that we separately developed, as well as split the tasks among the group in order to combine all of the elements at the end of the project. Furthermore, we were proud of the artificial intelligence use and being able to store the rankings in both the global mode and friends mode.
What we learned
What's next for Make it Match!
Built With
- css
- firebase
- html
- image-recognition
- javascript
- json
- machine-learning
- nosql
- transfer-learning
Log in or sign up for Devpost to join the conversation.