As university students and soon to be graduates, we understand the financial strain that comes along with being a student, especially in terms of commuting. Carpooling has been a long existing method of getting to a desired destination, but there are very few driving platforms that make the experience better for the carpool driver. Additionally, by encouraging more people to choose carpooling at a way of commuting, we hope to work towards more sustainable cities.

What it does

FaceLyft is a web app that includes features that allow drivers to lock or unlock their car from their device, as well as request payments from riders through facial recognition. Facial recognition is also implemented as an account verification to make signing into your account secure yet effortless.

How we built it

We used IBM Watson Visual Recognition as a way to recognize users from a live image; After which they can request money from riders in the carpool by taking a picture of them and calling our API that leverages the Interac E-transfer API. We utilized Firebase from the Google Cloud Platform and the SmartCar API to control the car. We built our own API using stdlib which collects information from Interac, IBM Watson, Firebase and SmartCar API's.

Challenges we ran into

IBM Facial Recognition software isn't quite perfect and doesn't always accurately recognize the person in the images we send it. There were also many challenges that came up as we began to integrate several APIs together to build our own API in Standard Library. This was particularly tough when considering the flow for authenticating the SmartCar as it required a redirection of the url.

Accomplishments that we're proud of

We successfully got all of our APIs to work together! (SmartCar API, Firebase, Watson, StdLib,Google Maps, and our own Standard Library layer). Other tough feats we accomplished was the entire webcam to image to API flow that wasn't trivial to design or implement.

What's next for FaceLyft

While creating FaceLyft, we created a security API for requesting for payment via visual recognition. We believe that this API can be used in so many more scenarios than carpooling and hope we can expand this API into different user cases.

Share this project: