Literature Review on Emotion Recognition System

摘要

Emotion plays a significant role in human beings daily lives. Humans can easily sense a person's emotions. But in some cases devices need to sense people's emotions. Machine learning is a sub-part of artificial intelligence that produces robots handling tasks like us. Emotion recognition is a small module that can be easily achieved by machines using machine learning algorithms. This paper describes the various algorithms used to recognize the facial expressions of a person such as happy, angry, sad, disgust, neutral, fear. Gabor filters and Local Binary Pattern Operators (LBP) are discussed for the process of feature extraction. Different types of classification algorithms such as Support Vector Machines, K-Nearest Neighbors are discussed. The training of the image data is carried by comparing various neural networks including Attentional Neural Network, Convolutional neural network, shallow neural network etc.