-
-
It's obviously negative.
Inspiration
Wanted a faster way to submit LFT results
What it does
Asks for user camera/image input to speed up results and verify test outcomes
How we built it
Get a bunch of friends at home to spam us LFT test photos as our dataset – then cropped and sifted for quality. Then used Lobe.ai to create a CNN, export as tensorflow.js and shove it into a React app. Send to Netlify, done.
Challenges we ran into
Tweaking model. Inaccurate results for null tests so leave for later implementation to improve on
Accomplishments that we're proud of
6 hours finished before deadline, no sleep
What we learned
Powered by Cisco™
What's next for LFT Neural Network Results Detection
Breakfast!
ps:
our model was working fine up until right before we submitted it. source: trust me bro
Built With
- lobe.ai
- netlify
- react
- tensorflow
