NOTE: THERE ARE 2 REPOS FOR THE PROJECT, ONE FOR PROCESSES RUN LOCALLY AND ONE FOR BUILDING AND RUNNING AN ENVIRONMENT ON HEROKU (linked here: https://github.com/SeanCena/slohacks20-heroku)
What it does
We have a program to sift through video feeds and, using face detection and ML
The project was built mostly with Python and various libraries, but incorporates a Firebase database for storing compiled lists of missing persons and related metadata and a Heroku server for running facial recognition models over the cloud. The project uses Google's Cloud Vision API for detecting faces from incoming footage before passing it on to the facial recognition models. Additionally, we built webcrawlers (for both static and dynamic web pages) to gather datasets to both train our vision models, and to detect real missing persons from around the world
Challenges I ran into
Some databases were either inconsistent with the amount of associated data they keep around each missing persons case, or were outright misleading on how many datapoints they have (cough, cough, INTERPOL)
Log in or sign up for Devpost to join the conversation.