A Construction Drawing Primitive Extraction System Based on CNN
| Autor: | Bin Wu, Song Houhou, Zhu Liangsheng, Quanyin Zhu, Hu Lingyu, Bian Wenwen, Feng Wanli |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
021110 strategic
defence & security studies Computer science business.industry Feature extraction Principal component analysis algorithm 0211 other engineering and technologies 02 engineering and technology Convolutional neural network Matrix (mathematics) Construction industry Dimension (vector space) 020204 information systems 0202 electrical engineering electronic engineering information engineering Feature (machine learning) Extraction (military) Artificial intelligence business |
| Zdroj: | 2019 18th International Symposium on Distributed Computing and Applications for Business Engineering and Science (DCABES). |
| DOI: | 10.1109/dcabes48411.2019.00010 |
| Popis: | In order to solve problem of huge manpower demand in the construction industry and heavy drawing work, a construction primitive extraction system was designed. This system relieved pressure of the artificial building review and improved the reviewing efficiency of the reviewing experts. The system uses a Convolutional Neural Network (CNN) to extract drawing primitives using the Visual Geometry Group (VGG) network training model. Principal component analysis algorithm improves feature comparison efficiency by reducing the dimension of the matrix. Experiments show that the practicability of the system can meet the requirements of general construction primitive extraction. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|