Emotion Recognition from Micro-Expressions using Logistic Regression and Stochastic Gradient Descent

Micro level expression recognition on the human face plays a vital role in nowadays research. A mechanism to automatically identify and extract semantic events is desperately needed. Micro expression shows real human emotion in a fraction of a second. Understanding expression from video and images based on identifying human emotion is used in many applications like crime detection, lie detector, and depression analysis in the medical field. To develop humanoid robots, Analysis of the physical condition of people and a variety of systems that involve interactions between persons and electronic devices such as human-computer interfaces are needed to implement humanoid robots. It is a required research domain for the following applications like social psychology, security and human-computer interaction.Micro expression detection mainly focuses on features of human facial parts. Human expression can be detected in different ways. In this paper, features extracted from micro expressions are used to identify different human activities. Before extracting features from micro expression, pre-processing the image is needed to identify the best features that are used to improve the accuracy of recognition and reduce the time needed for human emotion recognition. The extracted features are used by a combination of logistic regression and stochastic gradient descent algorithm to classify human emotion. CASME dataset is used to measure the performance of proposed work.