The story behind the code. We came in with no idea what we were going to develop...we hadn't even formed a team. After attending the workshop held by Clarifai about machine learning with their API. We just had to use this technology to make something amazing. After studying it and seeing potential for its facial recognition, we had the idea to make this app.
What it does
Our app, Twinder, analyzes an inputted picture by the user and searches among other users to find potential lookalikes. It returns the user's lookalike potentially allowing the user to find them and form a clone army!
How we built it
This application utilizes the Clarifai API to analyze inputted images and tag the photo with words that describe the image. Using this, we cross analyzed these tags with other tags in our database and narrowed down to find the lookalike of the user. Our algorithm also calculates the percentage likeness of the users to give the user an idea of how accurate the hack is!
Challenges we ran into
We came across many issues while attempting to build this hack. We changed languages multiple times trying to find a way to implement both back end and front end development in such a way so that everybody could contribute to the project. We finally came down to using PHP and jQuery and the hack came out to be successful.
Accomplishments that we're proud of
This project was as a whole an accomplishment we are all proud of. We came in with no ideas for a hack, with different skill sets that we thought would mismatch, and no plans to be a team. Rather than being a setback, all these things came together and everybody's different experiences helped us advance this hack to the next level.
What we learned
This hackathon was a great learning experience for us, we learned to work as a team and a few new languages alone the way. PHP was a brand new language for half of us and it turned out to be something we plan to implement in future projects.
What's next for Twinder
The future of Twinder is vast. We are hoping that with each advancement of Clarifai's API, we can further improve the reliability and accuracy of Twinder's facial comparison. We also want to increase our database to allow users to compare to celebrities, among Facebook friends, and as many more inputs as we could possibly find. Other possible improvements we were hoping to make in the future is to allow the users to communicate and meet up once they find each other. The domain we purchased for the domain.com challenge was whyuhefftobemad.com, but domain.com did not process it on time