LCA-GCS: Large City Architecture - Generated Cityscape Set

The Large City Architecture - Generated Cityscape dataset (LCA-GCS) is a comprehensive collection of 1,060,166 AI-generated images representing architectural features of 5,856 global cities. Created using advanced diffusion models, this dataset offers over 200 samples of various architectural types for each city with a population exceeding 100,000. LCA-GCS aims to facilitate comparative analysis, synthesis, and learning from AI-represented architecture, providing a unique resource for researchers, urban planners, and AI developers interested in global architectural trends and urban design.

Dataset Description

The LCA-GCS-set represents cities by situating architectural features from details to interiors to urban forms uncommon for existing databases in a flat ontology that allows to search and compare architectural representations in various ways. We hope that the LCA-GCS set benefits future research that could identify recurring spatial configurations, material palettes, and design elements characteristic of specific regions, cultures, or building typologies. Building on it, the synthetic dataset can contribute to e.g., training site-specific models, identifying commonalities, or extracting features that are difficult to access in existing data.

Methodology

Prompts: "Type in City." Where type iterates through the list of types below, and city iterates through the list of cities above 100,000 inhabitants listed at https://simplemaps.com/data/world-cities

Types: Apartment, Apartment Building, Arch, Atrium, Auditorium, Balcony, Bathhouse, Boulevard, Breezeways, Bridge, Brise Soleil, Bungalow, Bus Stop, Cafe, Campus, Canopies, Cityscape, Column, Communal Living Space, Courtyard, Door, Eaves, Enfilade, Entrance, Farm, Flat, Foyer, Forest, Garden, Hall, Habitat, House, Housing, Indoor Market, Indoor Plaza, Kitchen, Kiosk, Large Building, Large House, Living Room, Loft, Loggia, Lobby, Office, Park, Patio, Photo of the City, Piazza, Place, Playhouse, Porch, Rondavel, Roof, Roof Garden, Roof Terrace, Room, Row House, Semi-Detached House, Siedlung, Square, Staircase, Street, Terrace, Tower, Townhouse, Tree, Urban Block, Urban Forest, Urban House, Vegetation, Veranda, Vertical Garden, Villa, Window, Workshop Space.

Technical Details

Generated with Stability SD2.1, SDXL 1.0, SDXL-Turbo, SD3 using random seeds. For parameters per image, please see the metadata.csv

Citation

If you use the LCA dataset in your research, please cite it as follows: Koehler, D. (2024). Large City Architecture - Generated Cityscape Set (LCA-GCS): A Synthetic Dataset of Global City Architecture [Data set]. Huggingface. https://huggingface.co/datasets/Punktiert/LCA-GCS

Disclaimer of Warranties and Liability

Accuracy and Reliability The webpage "Large City Architecture" and its authors provide the data 'as is' and do not guarantee the accuracy, completeness, or usefulness of the data. No warranties are provided.

Limitation of Liability

To the fullest extent permitted by law, in no event will LCA-GCS or its authors be liable for any claims, damages, or other liabilities, whether in an action of contract, tort, or otherwise, arising from, out of, or in connection with the data or the use or other dealings in the data.

License

This dataset is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. For more details, visit the Creative Commons website.

Usage

To use this dataset locally or in your own projects, you can use the WebDataset library and ensure that your code supports the .webp format. To load this dataset with WebDataset:

import webdataset as wds
dataset = wds.WebDataset("path/to/tarfiles/{00000..00500}.tar").decode("pil")

for sample in dataset:
    img = sample['webp']  # Handle webp files