Abstract
Research on converting 2D raster drawings into 3D vector data has a long history in the field of pattern recognition. Prior to the achievement of machine learning, existing studies were based on heuristics and rules. In recent years, there have been several studies employing deep learning, but a great effort was required to secure a large amount of data for learning. In this study, to overcome these limitations, we used 3DPlanNet Ensemble methods incorporating rule-based heuristic methods to learn with only a small amount of data (30 floor plan images). Experimentally, this method produced a wall accuracy of more than 95% and an object accuracy similar to that of a previous study using a large amount of learning data. In addition, 2D drawings without dimension information were converted into ground truth sizes with an accuracy of 97% or more, and structural data in the form of 3D models in which layers were divided for each object, such as walls, doors, windows, and rooms, were created. Using the 3DPlanNet Ensemble proposed in this study, we generated 110,000 3D vector data with a wall accuracy of 95% or more from 2D raster drawings end to end.
Highlights
The history of architecture and architectural drawings is so long that it cannot be separated from human culture
Until the second half of the 20th century, when the spread of PCs made the use of computer-aided design (CAD) software [1] possible, architectural drawings were completed on paper and archived only on paper or as scanned raster images
We propose an ensemble method to convert a 2D floor plan image into a 3D model with similar accuracy to [16], even with a small amount of training data
Summary
The history of architecture and architectural drawings is so long that it cannot be separated from human culture. Converting raster images into vector-based drawings is a challenge for researchers, and, numerous studies have been conducted. Since an architectural plan represents the shape of a horizontal section at a specific height, its three-dimensional shape must be predicted, for example, a wall that is covered by the door/window image. These various objects must be classified and extracted, and their sizes should be meaningful. In order to overcome this limitation, we conducted learning with only a small amount of training data (30 floor plan images), and proposed an end-to-end 3DPlanNet Ensemble model that incorporates a complementary rule-based heuristic method. In order to service to the actual system [23], the insufficient parts were corrected manually, and a questionnaire survey on the performance of the 3DPlanNet Ensemble methods was conducted with the corresponding modifiers
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
