Enhancing the Interior Design Process with Crowdsourced Furnishing Pairing Recommendations

Abstract

Background Focusing on the importance of style and color in interior design, the necessity of comprehending the attributes of design elements is emphasized. While previous research lacks explicit rules for furnishing pairing recommendations specific to interior design styles, our study aims to address this gap through dataset analysis and machine learning techniques. The goal is to enhance designer-consumer communication in the interior design process, facilitating better-informed design choices. Methods Our methodology involved the analysis of a substantial dataset containing 24, 184 living room images sourced from Today’s House, an online home furnishing platform. We integrated data reflecting crowd preferences and applied machine learning for object detection and color extraction, converting visual information into quantifiable data. Additionally, we customized association rule mining to reflect crowd preferences, aiming to generate furnishing pairing rules specific to various interior design styles. We validated the effectiveness and practicality of these rules and our methods through expert interviews. Results The generated rules were organized based on three criteria: Adjusted-Support, Adjusted-Confidence, and Adjusted-Lift. We presented results that scored higher in these metrics, accompanied by illustrative image cases. These rules, having been validated through interviews with design experts, aid customers in making informed decisions and enhance designer-consumer communication and collaboration. Conclusions Our research presents a novel framework that enriches the interior design process through data-driven insights. The study’s contributions are threefold. First, we develop an Adjusted Association Rule Mining method for interior-style analysis and furnishing recommendations. Second, we demonstrate how different metrics like Adjusted-Support, Adjusted-Lift, and Adjusted-Confidence can be used to interpret and apply furnishing pairing rules effectively. Third, expert interviews confirm the utility of the rules in enhancing consumer decision-making and facilitating a more collaborative design process. We emphasize the importance of adaptable and predictive analytical methods in interior design and potentially other recommendation-based fields.