Neural Network based Recognition of an Emotion using PCA of Leg Postures of a Human Being

Abstract

This paper presents the recognition of an emotion of a human being based on Leg Postures created by an
individual human. Various leg postures related to an emotion are generated by different positions of legs viz. Sitting,
Standing and Walking. The image set of all these postures is prepared and analyzed by certain mathematical techniques
and image processing tools. Here, the images are classified into basic seven emotions Neutral, Happy, Sad, Fear, Anger,
Surprise, and Disgust by Neural Network (NN) which is applied on the eigen features of these images, obtained by
Principle Component Analysis (PCA).