Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia · 2019 · 12 citations · 0 references
ArchitectureDesignInterior DesignComputer VisionDesign AutomationArchitectural Design AutomationEngineering Design AutomationObject RecognitionPattern RecognitionAutomated Image AnalysisInterior Photography
This paper explores automated recognition of elements in interior design pictures for an intelligent design reference management system. Precedent design references have a significant role to help architects, designer and even clients in general architecture design process. Pictures are one of the representation that could exactly show a kind of design idea and knowledge. Due to the velocity, variety and volume of reference pictures data with growth of references platform, it is hard and time-consuming to handle the data with current manual way. To solve this problem , this paper depicts a deep learning-based approach to figuring out design elements and recognizing the design feature of them on the interior pictures using faster-RCNN and CNN algorithms. The targets are the residential furniture such as a table and a seating. Through proposed application, input pictures can automatically have tagging data as follows; seating1(type: sofa, seating capacity: two-seaters, design style: classic)