- Python - The code is designed to work with Python 2.7 and Python 3.5 or greater.
- NumPy - The randomized routines used for stochastic simulations are implemented using NumPy.
- TensorFlow - The core open source library to help you develop and train ML models. Get started quickly by running Colab notebooks directly in your browser.
- Pandas - pandas is a fast, powerful, flexible and easy to use open source data analysis and manipulation tool, built on top of the Python programming language.
To collect training and test data, run data_record.py. Various parameters are available in the file. The record data will be store into a folder call data
optional arguments:
--train Record training data or testing data. 1 for training data. 0 for testing data.
--store_method Choose the what kinds of store method you want to use. 1 for record the facial landmarks data. 0 for facial image
--expression_ID The data expression ID tag.
--show_landmarks If you want to show the landmark points in the camera screen.
--relative_coordinates Take the first landmark point as base point. Calculate the position of the rest of points based on the origin.
python landmark_train.py
python image_train.pypython real_time_tracking.py