earticle

영어

The representation of users' intends in the early design stage is critical to practitioners including architects. The development of an apparatus to support users' decision making by the representation of their intends efficiently has been conducted. This paper presents the framework development to automatically create Building Information Models from sketched data using deep learning technique. The methodology for the framework development is to define processes, the systems required in the framework and the data to develop the systems, and to implement the systems. The framework development consists of the following steps: the process model development, the framework implementation, and the validation of the framework. The process model development is to identify the activities to represent the process of the framework. The framework implemention is to develop the system interfaces and user interfaces which converts the sketched data from users into a Building Information Model. The validation of the framework investigates whether the framework can accurately create Building Information Models from the sketched data. The findings present that the framework can automatically translate sketched strokes in the sketched data into building components and accurately create in BIM models.