Facial expression recognition using Deep Convolutional Neural Network techniques
Ανίχνευση εκφράσεων προσώπου με χρήση Βαθιών Συνελικτικών Νευρωνικών Δικτύων
View/
Keywords
Convolutional neural network ; CNN ; Facial expression recognition ; FER ; FER-2013 ; VGG ; Machine learning ; ICML 2013 ; Deep learning ; Supervised learningAbstract
This dissertation constitutes a study of the classification accuracy of our model on the given dataset. Our supervised learning model uses a feed-forward neural network which we train with the backpropagation algorithm. The models we focus on are VGG-like Convolutional Neural Networks, trained and evaluated on the FER-2013 dataset. The algorithm classifies on seven emotion categories (Anger, Disgust, Fear, Happiness, Sadness, Surprise, and Neutral). We implemented this model with Python code on the CPU. All of our classifiers were implemented in Keras neural network API using TensorFlow backend. Our goal is to improve the classification accuracy on our dataset by choosing different architectures and by optimizing the model hyperparameters.
Postgraduate Studies Programme
Προηγμένα Συστήματα ΠληροφορικήςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
83Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού 4.0 Διεθνές
Αναφορά Δημιουργού 4.0 Διεθνές