The need for N95 masks in this crisis.
What it does
Places N95 masks over faces in an image, defending them against coronavirus.
How I built it
The script to add the mask uses dlib (ML library) and open-cv (Computer Vision library) to find the left cheek and tip of the chin of faces in an image. From there, we resize a png of an N95 mask and place it on top of each face.
The script lives in the back-end api made from flask. Since the computer vision data was fairly large, we had to dockerize the back-end component and run it on a Linux VM by AWS. The front-end is a React website.
Challenges I ran into
Running it in AWS EC2 on a Linux instance. Handling images as base 64 byte arrays
Accomplishments that I'm proud of
The domain is up and running. The Docker image is bulky but still running working.
What I learned
How to handle images as byte arrays How to run docker images on Linux containers provided by Amazon
What's next for 95Defender
Support for more image extensions Improve the UI
Try It out
amazon-web-services, dlib, docker, flask, node.js, open-cv, python, react.js