-
Detect and classify faces
-
face detection and classification with metadata
Inspiration
Advertisements on the streets are still so 20th century like. So it's time for change!
What it does
Based on the camera (right now camera on your mac, in the future camera on the add banner) it detects your face from the database. If you're not in the database, then it performs just basic segmentation, like your gender, age, race and clothing style.
How we built it
We've built kick ass backend using the state of the art Deep Learning methods, mainly using Keras and openCV libraries. All these functions are wrapped and served by the very simple Flask server (just for the demonstration purposes).
Frontend was built in AngularJS2.
Challenges we ran into
Not enough training data for the Deep Learning model. Way too short time for model training!
Accomplishments that we're proud of
Very accurate face detector, in very short time trained classifier.
What we learned
Better form the team before the hackathon including all the roles that you might need.
What's next for hackprague-face-recognition
Who knows, we let this open.
Built With
- javascript
- keras
- python
- tensorflow
Log in or sign up for Devpost to join the conversation.