Emotion Classification Comparison in Convolutional Neural Networks and Fuzzy Logics for Service Robotic Applications

Abstract

Emotions are a powerful indicator that can positively and negatively affect the long-term health of humans. Emotions are affected by many aspects of our daily lives, including decision-making, reasoning, and physical well-being. Additionally, emotions play a critical role in human interaction. Emotion recognition is a developing area in the development of service robotics system algorithms. This paper compares intelligent systems with convolutional neural networks and fuzzy logic for emotion recognition applications. Two system scenarios have been discussed separately to check the detection accuracy of facial expression recognition in image classification and speech signal recognition. The results reveal that convolution neural networks perform emotion recognition activities with higher accuracy than fuzzy logic applications. These experimental results can be used for future research applications to develop intelligent systems regarding emotional identification for service robotics. The developed convolutional neural networks applications have greater than 95% accuracy in accurate emotion detection.