Building Identification via Adjacency Graph Embeddings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods face challenges in effectively capturing and representing building interior information, including identifying buildings that match specific criteria, and using this information for navigation purposes without physically being present, as floor plans can be difficult to construct, maintain, and visualize accurately.
Innovation Solution
The use of computing devices to generate and compare hierarchical vector-based embeddings of building floor plans and attributes, allowing for automated identification of matching buildings by creating adjacency graphs and vector embeddings from images, enabling virtual navigation and personalized interactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional floor plans are used to represent building interiors, then some layout information can be provided, but the floor plans are difficult to construct, maintain, accurately scale, and visualize
Solution Approach 1:
The patent uses computer-generated images as copies of building interiors instead of traditional floor plans. These images are automatically generated from 3D building models, providing accurate visual representations without requiring manual construction or maintenance like traditional floor plans. The images capture layout, dimensions, and interior details while being easily updated when building changes occur.
Solution Approach 2:
The patent replaces the mechanical process of manually creating and maintaining floor plans with an automated computer-based system. The system automatically generates visual representations from digital building models, eliminating the need for manual drafting, scaling, and updating operations that characterize traditional floor plan creation.
2Loss of time
If physical travel to buildings is avoided, then time and resources are saved, but effective capture and representation of building interior information becomes difficult
Solution Approach 1:
The patent creates digital copies of building interiors through computer-generated images that can be viewed remotely. These images provide comprehensive interior information including layout, dimensions, and visual details, allowing users to explore building interiors without physical presence. The copies maintain accuracy and detail equivalent to or exceeding what could be observed in person.
Solution Approach 2:
The patent creates a universal digital representation system that serves multiple functions: it provides interior visualization, enables building identification and matching, supports navigation, and allows remote exploration. This single digital model replaces multiple physical activities including site visits, manual measurements, and physical inspections.
3Productivity
If automated building identification is implemented, then building matching efficiency is improved, but computational power and time requirements increase
Solution Approach 1:
The patent extracts and compares specific key features from building images and floor plans rather than processing entire datasets. The system identifies and compares essential characteristics such as layout patterns, room configurations, and structural features, reducing computational requirements while maintaining identification accuracy. This selective feature extraction enables efficient building matching with reduced processing power.
Data Source
AI summary
Techniques are described for using computing devices to perform automated operations for determining matching buildings and using corresponding information in further automated manners, such as by determining buildings having similarities to other indicated buildings based at least in part on building floor plans and/or other attributes, and by automatically generating descriptions of the similarities. In some situations, such matching building determinations are based on generating and using adjacency graphs for floor plans that represent inter-connections between rooms and other attributes of the buildings, and further generating and comparing vector-based embeddings that concisely represent the information of the adjacency graphs for use in performing inter-building comparisons and generating the descriptions. Information about such determined buildings may be used in various automated manners, including for controlling device navigation (e.g., autonomous vehicles), for display on client devices in corresponding graphical user interfaces, for further analysis to identify shared and/or aggregate characteristics, etc.