Blending Designer Insights With Predicted Crowd-Based Color Compatibility in Interior Color Design

Abstract

Harmonious color combinations define the quality and atmosphere of interior spaces. Designers refine schemes through iteration to meet client requirements, often relying on existing color design tools. However, these tools have two key limitations: they do not account for colors applied to specific objects, nor do they reflect aesthetic preferences. To address this, we developed a vision-language augmented Image Color Aesthetic Assessment model that predicts color compatibility for objects in interior design images. Moreover, this model powers the crowd-based Image Color Combination Evaluation system, enabling designers to prototype, evaluate, and generate new color combinations in real time. A user study with 16 designers revealed that 81% found the predicted color compatibility scores aligned with aesthetic principles with which they were familiar, aiding them in integrating these scores into their designs. The proposed system helps designers explore harmonious color combinations, avoid personal bias, and foster trust by leveraging predicted crowd-based color compatibility.