Skip to content
No description, website, or topics provided.
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
image_data
known_people
unknown_people
.gitignore
README.md
cam.jpg
face_recognition_video.py
face_recognition_video_emo.py

README.md

power_to_the_blind

MakeMIT Hackathon Submission

Problem Statement:

People with visual impairment may have trouble identifying people in their immediate vicinity. We built a Raspberry Pi-equipped hat that can feed people's names and relative locations into a user's earphones.

Details:

We trained a Convolutional Neural Network on a custom dataset of people's faces. The python-based immage classification model is then ported into a Raspberry Pi. We connected a mini camera to the Pi and ran the classification model. When a face is detected, the name of the person is read aloud out through the Pi's headpone jack, along with their relative position (left, middle, or right of image frame). The Pi setup is elegantly and discreetly attached to a baseball cap.

Team Members:

William Rodriguez, Dimitrios Koutentakis, Leandra Zimmerman

You can’t perform that action at this time.