In this project I have used convolution neural network to recognize human emotions from the their faces.
This dataset contains about 14.3k grayscale images equally distributed into 6 distinct types- Anger, Disgust, Fear, Happiness, Sadness, Surprise. There are 11,475 training images, 1433 validation images, and 1438 testing images.
- Conv2D, Max-pooling and Dropout layers.
- Fully connected dense layers for final classification.
evaluated_accuracy = 94.5%