Automatic analysis and sketch-based retrieval of architectural floor plans
Sheraz Ahmed, Markus Weber, Marcus Liwicki, Christoph Langenhan, Andreas Dengel, Frank Petzold
Article

Abstract:

The contribution of this article is twofold. First, we propose a sketch-based system, namely the a.SCatch system, for querying a floor plan repository. Second, for generating such a repository, a novel complete system for floor plan analysis is presented. The latter system extracts the semantics from existing floor plans. We introduce novel preprocessing methods, e.g., the differentiation between thick, medium, and thin lines and the removal of components outside the convex hull of the outer walls. Especially, the use of the convex hull increases the performance of the final system. The a.SCatch system enables the user to easily access knowledge from past projects. The user searches for semantically similar floor plans just by drawing parts of the new plan. An algorithm extracts the semantic structure sketched by the architect on DFKI’s Touch & Write table. Finally, the extracted structures are compared using the graph matching, and the most similar one is retrieved. In our experiments for floor plan analysis on a reference data set we compare our approach to other approaches available in the literature. We show that our floor plan analysis system outperforms previous systems. Also the performance of the sketch recognition system is quite high. Overall, the performance of the floor plan analysis, as well as of the retrieval, is already good for the use in practice.

Automatic analysis and sketch-based retrieval of architectural floor plans
(Hrsg.)
Article

Abstract:

The contribution of this article is twofold. First, we propose a sketch-based system, namely the a.SCatch system, for querying a floor plan repository. Second, for generating such a repository, a novel complete system for floor plan analysis is presented. The latter system extracts the semantics from existing floor plans. We introduce novel preprocessing methods, e.g., the differentiation between thick, medium, and thin lines and the removal of components outside the convex hull of the outer walls. Especially, the use of the convex hull increases the performance of the final system. The a.SCatch system enables the user to easily access knowledge from past projects. The user searches for semantically similar floor plans just by drawing parts of the new plan. An algorithm extracts the semantic structure sketched by the architect on DFKI’s Touch & Write table. Finally, the extracted structures are compared using the graph matching, and the most similar one is retrieved. In our experiments for floor plan analysis on a reference data set we compare our approach to other approaches available in the literature. We show that our floor plan analysis system outperforms previous systems. Also the performance of the sketch recognition system is quite high. Overall, the performance of the floor plan analysis, as well as of the retrieval, is already good for the use in practice.