MaskAlert

maskalert.tech

Inspiration

Many shop owners these days require that you wear a mask, but not everyone follows this requirement. Our product would alert the shop owner whenever someone who is approaching is not wearing a mask.

What it does

This product displays a live feed onto a webpage for the shop owner to view. When a customer approaches, the webpage will alert the shop owner if the approaching patron is not wearing a mask. Due to time constraints, our website requires the push of a button to test if there the person approaching is wearing a mask; however, a fully implemented project would automatically detect a person approaching and alert the shop keeper.

How we built it

We used Google's AutoML Vision API to build, train, and deploy a machine learning model that detects masks. We used Flask for the back-end and HTML and CSS for the front-end to develop our website and utilized OpenCV to integrate a live camera feed into the website.

Challenges we ran into

We had difficulty combining the two sides of the project (the website and the model). Both the mask detecting model and the live video feed website worked great separately, but we had trouble getting both working together.

Accomplishments that we're proud of

We are proud to have a final project that functions!

What we learned

It was the first time for all of us to work with GoogleCloud, coding for motion detection from a video, and tieing together the front-end and back-end of a website. There were many other things we learned about such as ideation, methods for collaborating while coding virtually, and trouble-shooting through using resources available to us.

What's next for MaskAlert

The project will be able to work better when we use more pictures to train it. Additionally, we could have the result of the mask detection model be printed onto the webpage rather than in the command prompt.

Note

Due to time constraints, we were unable to deploy to maskalert.tech. We have registered the domain name and plan to deploy following the hackathon.

Built With

Submitted to

WHACK 2020

Created by

I worked on the back end of the website, and implementing the live video feed. In addition to this, I contributed to getting the website to communicate with the mask detection model. I also created the logo.

Samuel Rose
I worked on two parts of the project. The first part is collecting images and setting up our Google Cloud AutoML Vision API to train for object detection, and the second is the creation of the video-pitch.

Jada Harrison
I worked on the front-end of the website using HTML and CSS. It was also my first time working on a project involving the back-end of a website and creating/training a model through Google Cloud.

Anandita Jha
Pursuing my interests in finance, technology, and entrepreneurship
I worked on the object detection portion of the project. This included integrating Google's Cloud AutoML Vision API with our website development.

Joseph Barberio