Key Words: Emotion recognition; emotion detection;intelligent agents; multimodal analyses;multilayered architecture.
Abstract. The paper presents different approaches for emotion detection and recognition. A system infrastructure for emotion recognition is suggested based on a multilayered multimodal architecture with intelligent agents. The agents have diverse goals, utilizes different modalities and employ various algorithms for emotion detection and recognition. Based on the results of the agents’ operation a neutral model of the multimodal input data is generated that is used for emotion recognition by detecting deformations with respect to the reconstructed model. The proposed system infrastructure allows autonomous operation, utilization of the most suitable recognition algorithms for certain input data and generation of neutral models for the examined object.