justcharles/access-control-project

Facial Recognition Access Control

A demo access-control system built on FaceNet embeddings (MTCNN detection + InceptionResnetV1 embedding) with an SVM / k-NN / small neural net ensemble for classifying a face as authorized, a specific enrolled person, or unauthorized.

• —Access Check tab: upload or capture a photo and get a grant/deny decision plus each model's individual prediction.
• —Enroll New Person tab: upload a few photos of a new person to add them as a recognized identity (retrains the SVM on the spot).

This is a personal/educational project, not a production security system.