Reputation-driven multimodal emotion recognition in wearable biosensor network

Emotion recognition is a process in which emotions are identified and recognized according to emotion-related bio signals. Wearable biosensor network expands the application of emotion recognition by measuring different emotion-related bio signals with wearable and portable hardware structures to meet specific needs in complicated measuring environment. This paper develops a multimodal emotion recognition method in wearable biosensor network. Reputation-driven Support Vector Machine (RSVM) classification algorithm is proposed to reduce the classification error caused by wearable sensor nodes. Reputations are figured out by similarity evaluation based on correlation calculation and are used for training sample selection and fuzzy membership degree determination. The experiment results indicate that this framework realizes reliable emotion recognition in wearable web-enable sensing environment and provides a solution to primary recognition and monitoring of emotional states and spiritual health.