-
Computer Vision
-
High accuracy facial landmarks
Inspiration
CANSOFCOM members are required to watch surveillance videos for extended period of time. Our program automates the process, freeing personnel of countless hours of manual observance.
What it does
Detect faces in realtime video stream through a webserver Recognize facial landmarks for each face on screen Count number of faces present in realtime and detect abnormality
How we built it
Read online tutorials about using opencv
Challenges we ran into
Installing dependencies to be compatible with other systems Training machine to recognize and remember seen faces
Accomplishments that we're proud of
Learning and coding in an unfamiliar programming language Python, opencv, dlib, and working with MongoDB in a short period of time
What we learned
Machine Learning takes a long time that we expected
What's next for Excubia
Using data in the database to identify and track individuals' occurrences and detect abnormality Not only track faces, but also body movements
Log in or sign up for Devpost to join the conversation.