VGG-16 and Optimized CNN for Emotion Classification

Abstract

This paper discusses the efficiency of using VGG 16 and Optimized CNN on emotion classification of human face dataset. Facial expression can be used as a communication medium between people to express a person feeling not only what it has shown on the outside also include the inner feeling, mental situation and perspective. In this paper, 5 basic emotions have been chosen to test the highly efficient models to elaborate the model complexity of the state-of-the-art model towards its accuracy and training efficiency of the proposed models against the state-of-the-art model such as VGG16 to achieve at least 65% accuracy.