Application of Convolutional Neural Network for Facial Expression Recognition

The recognition of people’s expression has been a very difficult task for computers from the time of its invention and still continues to pose a lot of challenges to the modern day generation of computers. To solve this problem, Convolutional Neural Network (CNN) is used which involves the application of preprocessing, feature extraction, training technique, and testing modules/methods to determine facial expression recognition. These methods were tested on the Oulu-CASIA VIS dataset [1]. The results obtained classified images of people’s facial expressions into six (6) distinct emotional classes, viz (anger, disgust, fear, happiness, sadness and surprise) showing an average accuracy of 98.99% and thus affirming that the application of the convolutional neural network (CNN) in facial expression recognition is efficient.