Stylization-Based Floor Plan Generation via Graph Convolutional Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current floor plan design tools require significant expertise and are challenging to use without extensive knowledge, limiting their accessibility for generating high-quality designs, especially in residential and non-residential facilities.
Innovation Solution
A stylization-based floor plan generation apparatus using a graph convolutional message passing network, space layout network analyzer, and cascaded alignment layer analyzer processes user inputs to generate realistic floor plans, allowing for intuitive design with limited knowledge and expertise, and enables transfer of styles between floor plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If CAD tools are used for floor plan design, then design quality can be improved, but the ease of operation deteriorates due to required expertise
Solution Approach 1:
The patent introduces an intermediary system comprising a layout graph generator, graph convolutional network, and floor plan generator that mediates between simple user inputs and complex CAD design outputs. This intermediary automatically processes user requirements into professional floor plan designs, eliminating the need for users to directly operate complex CAD tools while maintaining high design quality.
2Manufacturing precision
If expert architects perform floor plan design, then design quality improves, but productivity decreases due to manual iteration and feedback cycles
Solution Approach 1:
The system enables self-service floor plan generation by automatically processing user inputs through the layout graph generator and graph convolutional network to produce complete floor plan designs without requiring expert architects to manually sketch, iterate, and refine designs. The automated generation process significantly improves productivity while maintaining design quality through algorithmic optimization.
Solution Approach 2:
The patent replaces the mechanical process of manual floor plan design with an automated computational system. The graph convolutional network and generative model substitute the human expert's manual sketching and iteration process, transforming a labor-intensive mechanical process into an efficient automated computational workflow that maintains design quality while dramatically improving productivity.
3Adaptability or versatility
If complex designing tools are used, then design capability is improved, but device complexity increases making them challenging to utilize
Solution Approach 1:
The patent segments the complex design tool into distinct functional modules: a layout graph generator for processing user inputs, a graph convolutional network for spatial reasoning, and a floor plan generator for output generation. Each module handles a specific aspect of the design process, making the overall system more manageable and easier to use while maintaining comprehensive design capability.
Solution Approach 2:
The patent creates a universal floor plan generation system that can handle multiple types of design requirements through a single integrated platform. The graph convolutional network and generative model provide multi-functional capabilities to process various user inputs and generate diverse floor plan designs, eliminating the need for multiple specialized tools and simplifying the user experience.
Data Source
AI summary
In some examples, stylization-based floor plan generation may include receiving, for a floor plan that is to be generated, a layout graph for which user constraints are encoded as a plurality of room types. The user constraints may include spatial connections therebetween. Based on the layout graph, embedding vectors may be generated for each room type of the plurality of room types. Bounding boxes and segmentation masks may be determined for each room embedding from the layout graph, and based on an analysis of the embedding vectors for each room type of the plurality of room types. A space layout may be generated by combining the bounding boxes and the segmentation masks. A floor plan may be generated based on an analysis of the space layout and an input boundary feature map.