Comparative Analysis of Emotion Detection using CNN and Multilayer Perceptron
This paper talks about the application of feature extraction of facial expressions with the help of a combination of neural networks for the detection of different sets of emotion such as happy, sad, angry, fear, surprise, neutral etc. We humans are capable of generating hundreds of facial expressions when we communicate with each other that differ in levels of intensity, complexity, and meaning. This paper analyses the limitations of existing systems for emotion detection. Carrying out the experiment, we have achieved 93 percent accurate results with the help of convolution neural networks and it is easier and simpler than Emotion detection using a Multilayer Perceptron where 87 percent accuracy was achieved. Existing Human Emotion Recognition Systems using MLP have been compared with Emotion Recognition Systems using CNN. By this paper it was found out that neural network obtained better results for the system. Keywords - Emotions, Feature Extraction, Neural Network, Emotion Recognition, Emotion Detection, Facial Recognition, Convolutional Neural Network, Multilayer Perceptron.