This is the user Interface:
Features:
- Robust in anti-cheating measures
- 99.4% Accuracy in detection.
- Attendance with Time is stored into the database.
- Very Accurate in Different Light Conditions especially low-light
- Searches if the newly registering user already exists in the database or if his name or id are already present in the system.
- Ability to Safely remove a user from the system
- Confidence Score system to prevent false positives
Algorithm:
- Uses Dlib's Histogram of Oriented Gradients Technique for Face Detection and Orientation
- Uses FaceNet (Deep Learning Model) for Face Identification
- Uses SVM Classifier to query database rapidly (to match face)
- Eye Blink Detection for anti face-cheating
Implementation:
- Used Python Language
- OpenCV, dlib, face_recogntion Libraries
This code is deployment ready upon slight few changes
Future Work
- Will add more features which come accross my mind