The above script is my implementation of the following paper for facial expression recognition, which makes use of the CK+ database.
Jung, Heechul, Sihaeng Lee, Junho Yim, Sunjeong Park, and Junmo Kim. "Joint fine-tuning in deep neural networks for facial expression recognition." In Proceedings of the IEEE International Conference on Computer Vision, pp. 2983-2991. 2015.
The model reaches 96.2% subject-independent accuracy on joint fine-tuning.
fer2013 folder contains training script for fer2013 database. This model reaches an accuracy of 59.8% over random 10% of the dataset used as testing set.