A Data-Driven Framework for Predicting Emotional Responses to Plant Color in Healing Landscapes
International Journal of Healthcare Information Systems and Informatics · 2026-11-02 · 21 (1) · 1-17
DOI: 10.4018/ijhisi.426661 ↗Abstract
Traditional healing landscape design lacks quantitative analysis and objective prediction of plant color configuration. We propose a data-driven framework integrating computer vision with physiological feedback. The Color Configuration Index (CCI) is extracted from Hue, Saturation, Value (HSV) space and color entropy. Using virtual reality (VR) experimental data from 60 subjects (electroencephalography [EEG], heart rate variability [HRV], and Profile of Mood States [POMS]), we build a multimodal multilayer perceptron (MLP) emotion prediction model and employ structural equation modeling (SEM) with Gaussian fitting to reveal the mechanisms of color-driven recovery. Color entropy exhibits an inverted U-shaped dose–effect relationship with the fitted valence peak at Hopt = 1.85 and the fitted EEG α peak at Hopt = 1.75; the unified optimal interval is 1.75–1.85. Perceived biodiversity mediates about 70% of the CCI effect on psychological recovery. This work offers a parametric tool for precision healing design and decision support in health environment design.
EUVIMED