Comet Pass

• 

  This is the landing page of the Facial Authorizer, see our video demo link for more!

Inspiration

There have been many moments when our team members struggle to retrieve their comet card in order to enter a facility or sometimes they may have forgotten it at home. The student can also enter using their GET app, but then again the student needs to shuffle through their pockets to retrieve their phone. We decided to make the student experience more efficient by solving this issue.

What it does

All the student needs to do is look towards the camera. The Comet Pass Facial Authorizer uses a camera to detect a students face and checks if they are in the UTD database. If they are, then they are checked-in, if they are not, then they can either offer to enter using their comet card, or they will be sent home. The Facial Authorizer is very accurate using AWS software.

How we built it

The Facial Authorizer uses an AWS tool called "Rekognition" which offers a variety of features including face matching. We would then upload the UTD directory of student pictures inside of the S3 bucket, an AWS storage system. When the student gets in front of the camera and hits "check-in" it then snaps a picture and checks for a similar image in the AWS S3 bucket. We used Flask for the backend server and React JS for the front-end of the webapp.

Challenges we ran into

We were not familiar with the AWS platform, so that took a little bit of time to become familiarized with.

Accomplishments that we're proud of

Our Facial Authorizer is very accurate and can even recognize faces through low quality pictures. The check-in is also immediate and does not take time to process.

What we learned

We learned how to implement API's using the AWS system. We learned how to take snapshots of pictures and comparing it to a bucket of pictures(data).

What's next for Comet Pass

We would like to implement new features including checking-in students as they walk into a facility without having to stand in front of the camera and hit "check in". We will also flag individuals who do not pass the check-in so that they may be turned down.

Built With

• amazon-web-services
• flask
• javascript
• python
• react
• rekognition