Emotion Recognition from EEG Waves

This project was done under Professor Veeky Baths. The major idea here was to treat the Electroencephalographic Waves obtained from the brain as images, and pass them through a convolutional neural network to observe whether we could obtain any insights as to whether they could be useful for extracting emotions. We used the DEAP dataset for this purpose, and we tried to predict the Valence and Arousal values from the waves. They were each split into 4 classes based on the range of values where they each belonged to. We obtained an accuracy of 70% for Valence and 65% for Arousal.