Face Recognition Logger Using Tensorflow

This is an automated facial recognition logger for video security surveillance using TensorFlow implementation of the face recognizer described in the paper "FaceNet: A Unified Embedding for Face Recognition and Clustering".

Compatibility

The code is tested using TensorFlow 1.12 under Windows 10 with Python 3.6.

Installation

1. Import bioeye.sql
2. Configure the cameras in config.ini. Example:

cam0 = "https://root:ismart12@192.168.1.2/cgi-bin/currentpic.cgi"
log0 = "in_log"
cam1 = "https://root:ismart12@192.168.1.3/cgi-bin/currentpic.cgi"
log1 = "out_log"

3. Run bioeye.exe. All cropped faces will be stored in the cluster folder.
4. Run cluster.py to segregate alike faces.
5. Rename each folder in the train_img folder to identify each person.
6. Press Ctrl+T in bioeye.exe to train the images in the train_img folder.

Inspiration

The code is heavily inspired by the Facenet implementation.