Extracting Architectural Design Elements of CAD Data Using Deep Learning Algorithms

Abstract

In computer-aided design (CAD) data, architectural design elements usually exist in various forms, such as graphics, lines, and texts. Extracting these design elements efficiently and accurately from massive CAD data and applying them to architectural design practice is a difficult problem in architectural design. In this article, deep learning (DL) and fuzzy C clustering (FCM) algorithms in data mining (DM) are combined to extract architectural design elements from CAD data, predict the required architectural features, and provide support for the intelligent development of architectural design. By comparing the recall and accuracy, it is found that the algorithm can effectively identify the actual architectural features and has a high proportion of real cases in the samples predicted as positive cases. This shows that the algorithm can not only capture the architectural features but also effectively eliminate the interference factors and reduce the occurrence of false positives. This method improves the performance of extracting design elements and provides reliable technical support for intelligent architectural design and promotes cross-domain innovation and development.