The Application of AI-Assisted Color and Emotion Association Modeling in Healing Space Design

This study addresses the issues of high subjectivity and lack of quantitative basis in traditional color design for therapeutic spaces by proposing an AI-assisted method for modeling the relationship between color and emotion. First, multimodal emotion recognition technologies (such as facial expression analysis, speech, and semantic analysis) are used to extract users’ current emotional states. Then, a mapping relationship between color parameters (HSV/RGB) and the Valence-Arousal emotion model is constructed to achieve a quantitative conversion between "emotion and color." Finally, an intelligent color-matching system is developed and embedded into the spatial design process to assist designers in generating emotionally guided color schemes. The study adopts a psycho-physiological joint evaluation method, using tools such as the PANAS and SAM scales, as well as physiological indicators like EDA and HRV, to verify the model's effectiveness. The results show that the proposed method significantly improves emotion recognition accuracy and the generation of emotionally congruent color schemes, with high user satisfaction. This research provides AI-based support for personalized and scientific therapeutic space design, offering promising application prospects.