Machine Learning System for Architectural Renderings and BIM Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Building design is labor-intensive and expensive due to the high costs associated with creating multiple architectural renderings, with Star Architects' fees often accounting for up to 15% of the project budget, driven by the need for distinctive design aesthetics.
Innovation Solution
A machine learning system utilizing a Generative Adversarial Network (GAN) generates realistic building renderings and BIM data based on user-input surfaces and architectural constraints, reducing the need for manual design and lowering costs by automating the rendering process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multiple architectural renderings are created by hand or using computer-assisted design tools, then distinctive design aesthetics and insights are achieved, but labor costs and project expenses increase significantly
Solution Approach 1:
The patent replaces the mechanical system of manual architectural rendering with an automated machine learning system. The ML system processes simplified building depictions and automatically generates photorealistic renderings, substituting human architects' manual work with computational algorithms while maintaining design quality.
Solution Approach 2:
The patent creates multiple copies of building renderings by applying different architectural styles to the same building structure. The ML system can generate numerous style variations (e.g., Victorian, Modern, Colonial) from a single input, enabling designers to explore multiple aesthetic options without proportional increases in cost.
2Adaptability or versatility
If multiple architectural renderings are created to explore different design aesthetics, then design options are expanded, but time consumption and project duration increase
Solution Approach 1:
The patent enables rapid periodic generation of design variations by allowing users to systematically apply different architectural styles one after another. The ML system can quickly cycle through multiple style applications, generating a series of rendered options in sequential fashion, dramatically reducing the time required to explore design alternatives.
Solution Approach 2:
The patent performs preliminary processing of building structures by first creating simplified depictions or wireframes that capture essential geometric information. This preliminary representation is then used as input for rapid style application, allowing the system to prepare the foundation for multiple design explorations in advance.
3Manufacturing precision
If photorealistic building images are generated using traditional methods, then visual accuracy is achieved, but the process becomes labor-intensive and expensive
Solution Approach 1:
The patent replaces traditional manual or software-based rendering processes with a machine learning-based automated system. The ML model has been trained to generate photorealistic images directly from simplified inputs, substituting the complex mechanical processes of traditional rendering with intelligent computational generation.
Solution Approach 2:
The patent changes the key parameter from detailed geometric input to simplified building depiction as the primary input. By transforming the input representation and using trained ML models, the system achieves photorealistic outputs with significantly reduced computational complexity and faster processing times compared to traditional rendering methods.
Data Source
AI summary
Techniques are disclosed for using a computation engine executing a machine learning system to generate, according to constraints, renderings of a building or building information modeling (BIM) data for the building, wherein the constraints include at least one of an architectural style or a building constraint. In one example, an input device is configured to receive an input indicating one or more surfaces for the building and one or more constraints. A machine learning system executed by a computation engine is configured to apply a model, trained using images of buildings labeled with corresponding constraints for the buildings, to the one or more surfaces for the building to generate at least one of a rendering of the one or more surfaces for the building according to the constraints or BIM data for the building according to the constraints. Further, the machine learning system is configured to output the rendering or the BIM data.