Facial Landmarking

facial_landmarking.py

Completed

This is a Python script which utilizes the shape_predictor_81_face_landmarks.dat file to facially landmark 81 locations of our face. The dlib library is used to first detect the faces and the dat file is used for the landmarks.

An extra touch to the script was the use of counting how many blinks were made. This was done via the Ear-Aspect-Ratio(EAR) which makes use of the 6 coordinates surrounding the eyes. The following GIF gives a representation of the project.

Installation

Requirements

The following software libraries/frameworks should be installed:

• Python 3.3+ or Python 2
• imutils library
• datetime library
• time library
• dlib package
• cv2 package

  Installation Options

  The libraries and frameworks displayed above could be installed in the following direction:

pip3 install 'framework/library'

Usage

1. Download facial_landmarking.py.
2. Download shape_predictor_81_face_landmarks.dat
3. Go to your Python IDLE and insert:
   python3 facial_landmarking.py