EMOTION IDENTIFICATION IN VIDEO USING CONVOLUTIONAL NEURAL NETWORK

Abstract

Emotion analysis and identification in a video are based on visual features of the video clip to identify and classify the emotion in this clip. The analysis requires sufficient feature extraction from visual data. Video data usually in high dimension and require high-performance computing with an efficient algorithm, especially needed in real-time analysis. In this study, we propose a new feature extraction method suitable for emotion classification in video problem. We also propose a new classification model base on the improvement of modern deep learning is Convolutional Neural Network. The new proposals are verified on an experiment of video emotion classification and show the better performance compared with traditional machine learning.