Generative Adversarial Network for Architectural Plan Vectorization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing systems struggle to efficiently convert simple format architectural plans into machine-readable vectorized formats, as they are time-consuming and prone to errors, especially for untrained individuals and machines, due to variations in formats and notation styles.

Innovation Solution

A method using a trained machine learning model, specifically a generative adversarial network, is employed to input architectural plans, split them into layers, and undergo post-processing for contourization, room segmentation, and simplification to generate vectorized plans, which can be manually corrected.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If manual tracing of vector shapes is used to convert architectural plans, then the conversion accuracy is improved, but the time consumption increases significantly

Engineering Contradiction:

Improveconversion accuracyVSAvoidtime consumption

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs preliminary processing by detecting architectural features and generating initial vector shapes automatically before manual tracing, reducing the time required for manual correction while maintaining accuracy

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces the manual mechanical tracing process with an automated computer vision system that detects architectural features and generates vector shapes, significantly reducing time consumption while maintaining conversion accuracy

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If computer vision and image processing methods are used to convert architectural plans, then the processing speed is improved, but the error rate increases

Engineering Contradiction:

Improveprocessing speedVSAvoiderror rate

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system segments the architectural plan into distinct features (walls, windows, doors, text) and processes each separately with specialized detection algorithms, improving accuracy while maintaining processing speed

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary step where detected features are validated and corrected against architectural constraints and rules before final vector generation, reducing errors while maintaining processing efficiency

Inventive Principle:

Principle #24Intermediary (Mediator)

3Ease of manufacture

If simple format images are used for architectural plans, then the ease of creation is improved, but the machine readability decreases

Engineering Contradiction:

Improveease of creationVSAvoidmachine readability

Core Design Contradiction:

Ease of manufactureVSDifficulty of detecting and measuring

Solution Approach 1:

The patent replaces manual format conversion with an automated system that uses computer vision to detect and interpret architectural features directly from simple format images, converting them into machine-readable vector data without requiring manual intervention

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS20230214556A1Method and system for generating architectural vector maps

Publication Date: 2023.07.06 MAPPEDIN

AI summary

A method and associated system for generating vectorized architectural plans. The method includes inputting an architectural plan into a trained machine learning model, receiving a translated architectural plan as an output from the trained machine learning model, and post processing the translated architectural plan to generate a vectorized architectural plan.