Deep Learning Approach for Emotion Recognition from Human Body Movements with Feedforward Deep Convolution Neural Networks

Abstract

Nowadays, analysis of human body movements for emotion recognition is essential for social communication. Non-verbal communication methods like body movements, facial expression, gestures and eye movements are used in several applications. Among them emotion recognition from body movements has an advantage of recognize emotions of person from any camera view and also recognize emotions, if person is too far from camera. The body movements can strongly convey emotional states than other studies. In this paper, emotional state recognizes from full body motion patterns using feedforward deep convolution neural network architecture with different parameter. The proposed system can be evaluated by emotion dataset (University of YORK) with 15 types of emotions and GEMEP corpus dataset with 5 emotions. The experimental result showed the better recognition accuracy of the proposed system.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 80 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.