Alexandra Raibolt ( Lattes | E-mail )
This Jupyter Notebook shows step by step, the process of building a Convolutional Neural Network for Emotional Expression Recognition in Python using the TensorFlow framework.
In this example we use the Extended CK+ dataset.
Notice:
-
The LBCNN model proposed in this work was implemented in Python (version 2.7.12) using the TensorFlow framework (version 1.4.0) using a GPU based architecture, and might not work with other versions.
-
The directory where the datasets should stay is not available in GitHub, since it would violate the dataset rules.
- datetime
- scipy.stats
- sklearn.externals
- sklearn.metrics
- gzip
- itertools
- matplotlib
- numpy
- os
- tensorflow
- time
You can install missing dependencies with pip. And install TensorFlow via TensorFlow link.
- Install the dependencies;
- Run Jupyter Notebook in terminal to see the code in your browser.
-
Kanade, Takeo, Yingli Tian, and Jeffrey F. Cohn. "Comprehensive database for facial expression analysis." fg. IEEE, 2000.
-
Lucey, Patrick, et al. "The extended cohn-kanade dataset (ck+): A complete dataset for action unit and emotion-specified expression." Computer Vision and Pattern Recognition Workshops (CVPRW), 2010 IEEE Computer Society Conference on. IEEE, 2010.
Code released under the MIT license.