Kos1976/9-facades-doors-windows-50-commercial

Architectural Facades and Details — 50 Authentic European Photographs

📋 Dataset Summary

Curated collection of 50 high-resolution photographs documenting authentic European architectural details — facades, doors, windows, walls, arches, gates, balconies, and columns. From rustic Italian farmhouses and Tuscan alleyways to ornate stone archways, Gothic cathedral entrances, medieval iron gates, historic French academy portals, and charming countryside barn doors.

Each image includes comprehensive CSV metadata with 14 classification fields optimized for machine learning pipelines, generative AI training, computer vision benchmarks, and heritage preservation research.

What makes this dataset unique

• Authenticity: All photographs shot on-location across Europe — no stock imagery, no synthetic generation
• Real locations: Shot across Italy, Rome, Greece, Mediterranean regions, and Europe
• Authentic weathering: 27 of 50 images show genuine age, patina, and historical wear — not artificial distressing
• Material diversity: Predominantly stone (26), plus brick (13), plaster (8), wood (2), and iron (1)
• Architectural context: Elements framed by stone archways, brick facades, stucco walls, marble columns, wrought iron grilles, and cobblestone streets
• Rich metadata: ~42 descriptive keywords per image + structured classification fields
• No missing data: All 14 metadata fields are fully populated for every image

---

📊 Dataset Statistics

Metric	Value
Total images	50
Primary elements	Facade (30), Door (11), Arch (5), Window (2), Balcony (1), Gate (1)
Dominant material	Stone (26), Brick (13), Plaster (8), Wood (2), Iron (1)
Style distribution	Neoclassical (15), Baroque (7), Gothic (6), Mediterranean (6), Renaissance (5), Romanesque (3), Classical (2), plus Art Nouveau, Russian Revival, Vintage, Rustic, Modernist
Condition	Weathered (27), Restored (21), Authentic (2)
Historical eras	19th Century (23), Medieval (13), Baroque (3), Early 20th Century (3), Renaissance (2), Ancient (2), plus Late 19th Century, 21st Century, 17th Century, Mid 20th Century
Building types	Residential (19), Religious (14), Public (12), Commercial (4), Unknown (1)
Geographic coverage	European (36), Europe (3), Italy (2), Mediterranean (2), USA (2), Asia (1), Rome (2), Greece (1)
Color palette	Red (47), Orange (20), Yellow (5), Blue (5), Brown (4), Green (3), White (3), Neutral (2), Pink (1)
Keywords per image	~42 average (range: 30–46)
Description length	~328 characters average (range: 290–382)
Title length	~32 characters average (range: 24–43)
Format	JPEG (.jpg)
Metadata	Structured CSV with 14 fields

Detailed Element Distribution

Element	Count
Facade	30
Door	11
Arch	5
Window	2
Balcony	1
Gate	1

Detailed Style Distribution

Style	Count
Neoclassical	15
Baroque	7
Gothic	6
Mediterranean	6
Renaissance	5
Romanesque	3
Classical	2
Art Nouveau	1
Russian Revival	1
Vintage	1
Rustic	1
Modernist	1

Detailed Era Distribution

Era	Count
19th Century	23
Medieval	13
Baroque	3
Early 20th Century	3
Renaissance	2
Ancient	2
Late 19th Century	1
21st Century	1
17th Century	1
Mid 20th Century	1

Detailed Color Distribution

Color	Frequency
Red	47
Orange	20
Yellow	5
Blue	5
Brown	4
Green	3
White	3
Neutral	2
Pink	1

---

🏷️ Metadata Schema

Each image is accompanied by detailed metadata in photos_metadata.csv:

Field	Type	Description	Example Values
filename	string	Image file name	IMG (1).jpg · IMG (50).jpg
title	string	Commercial descriptive title (24–43 chars)	"Gothic Church Entrance Portal" · "Neoclassical Facade with Columns"
description	string	Detailed description (290–382 chars) with style, period, materials, mood	"The image showcases a Gothic-style church entrance with an intricately carved stone portal..."
keywords	string	30–46 comma-separated tags	facade, arch, church, romanesque, medieval, stone, columns, relief...
element	string	Primary architectural element	facade · door · window · arch · balcony · gate
building_type	string	Building type	residential · religious · public · commercial · unknown
material	string	Dominant construction material	stone · brick · plaster · wood · iron
style	string	Architectural style	neoclassical · baroque · gothic · mediterranean · renaissance · romanesque · classical · art_nouveau · modernist
condition	string	Physical state	weathered · restored · authentic
color	string	Dominant color palette (comma-separated)	red, orange · red, blue, yellow · red, green, orange, brown, white
era	string	Historical period	19th_century · medieval · baroque · early_20th_century · renaissance · ancient · 21st_century
location	string	Geographic region	european · italy · rome · mediterranean · greece · usa · asia
author	string	Photographer	Konstantin Anikin
copyright	string	Rights holder	(C) 2026 Konstantin Anikin

Data Splits

This is a single-split dataset. All 50 images are in the train split.

configs:
- config_name: default
  data_files:
  - split: train
    path: "photos_metadata.csv"

---

🎯 Recommended Use Cases

🏠 Interior Design AI

Training data for room generation models (RoomGPT, Interior AI, Collov) — real door textures, proportions, and styles for authentic interior visualization.

🏛️ Architectural Visualization

Reference library for 3D artists and architects working on heritage restoration, historical reconstructions, or European-style architectural projects.

🎨 Generative AI & Diffusion Models

• Interior design AI — training data for room generation (RoomGPT, Interior AI, Midjourney architectural styles)
• Architectural visualization — texture references and style anchors for 3D rendering pipelines
• Heritage preservation — authentic patina, weathering patterns, and material aging for restoration AI
• Image captioning — rich descriptions and keyword tags for vision-language model training

📐 Computer Vision Benchmarks

• Object detection — architectural element classification (facade, door, window, arch, balcony, gate)
• Material recognition — stone/brick/plaster/wood/iron texture classification
• Multi-label classification — color, style, era, condition, and building type prediction
• Style transfer — neoclassical/baroque/gothic/renaissance architectural style embeddings
• Color analysis — Mediterranean/European heritage color palette extraction

🧱 3D Texture Generation

Photogrammetry and NeRF training input for creating realistic PBR textures of weathered stone, aged brick, plaster walls, wood, and wrought iron.

📸 Stock Photography & Editorial

Curated visual reference for travel, architecture, and lifestyle editorial content.

🎮 Game Development

3D texture generation for historical/fantasy environments — medieval castles, Gothic cathedrals, rustic villages, ancient fortresses, neoclassical palaces.

---

🖼️ Sample Images

Title	Element	Material	Style	Condition	Era	Location
Historic Residential Facade with Bay Window	facade	plaster	art_nouveau	weathered	early_20th_century	european
Renaissance Building with Arches	facade	brick	renaissance	restored	renaissance	italy
Baroque Archway Entrance	arch	stone	baroque	weathered	19th_century	european
Gothic Church Entrance Portal	door	stone	gothic	weathered	medieval	european
Neoclassical Facade with Columns	facade	stone	neoclassical	restored	19th_century	european
Romanesque Church Portal Detail	door	stone	romanesque	weathered	medieval	european
Baroque Facade with Atlantes	facade	stone	baroque	restored	19th_century	european
Ancient Brick Dome Facade	facade	brick	rustic	weathered	ancient	europe
Modernist Residential Facade	facade	plaster	modernist	restored	mid_20th_century	greece
Colosseum Exterior Arches	arch	stone	classical	weathered	ancient	rome

(Full previews available in the image files)

---

📁 File Structure

architecture-facades/
├── IMG (1).jpg
├── IMG (2).jpg
├── ... (50 JPEG files total)
├── photos_metadata.csv    # Full structured metadata for all 50 images
├── README.md              # This file
├── COMMERCIAL_LICENSE.md  # Commercial licensing terms (detailed)
└── LICENSE.txt            # License agreement (summary)

---

📜 Licensing

© 2026 Konstantin Anikin. All rights reserved.

This dataset is available under a CC-BY-4.0 license for academic and research use, with a custom commercial license available for commercial applications.

Academic / Research Use (Free)

• Non-commercial use with mandatory attribution
• Citation required in all publications
• See LICENSE.txt for full terms

Commercial Use

For commercial use, bulk licensing, or custom dataset requests, please contact the author.

License Type	What's Included	Price
Academic / Research	Non-commercial use, citation required	Free (CC-BY-4.0)
Commercial — Startup	1 AI model, up to 100K users	€200
Commercial — Business	Unlimited models, unlimited users	€500
Exclusive	Full ownership, dataset removed from public	€1,500

Custom licensing available upon request. Bulk discounts for 3+ datasets.

📧 Contact: tehnomaster1973@gmail.com
Subject: Commercial License Request — Architectural Facades Dataset

See COMMERCIAL_LICENSE.md for complete commercial license terms.

---

📝 Citation

If you use this dataset in your research or project, please cite:

@dataset{konstantin-anikin_architecture-facades,
  author    = {Konstantin Anikin},
  title     = {Architectural Facades and Details},
  year      = {2026},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/konstantin-anikin/architecture-facades},
  keywords  = {architecture, facades, doors, windows, walls, arches, heritage, europe, computer vision, generative AI}
}

---

👤 Author

Konstantin Anikin — photographer and mechanic based in Italy

• 🔗 Shutterstock: [https://www.shutterstock.com/g/kos1976]
• 🔗 Amazon KDP: [https://www.amazon.com/author/darkacademia]
• 📸 Contributing photographer on Shutterstock, Adobe Stock, Dreamstime, Depositphotos, 123RF
• 🏗️ Specializing in architectural photography, heritage documentation, and industrial details
• 🤖 Building curated training datasets for generative AI and computer vision

Passionate about documenting European architectural heritage through authentic, unfiltered photography. This dataset represents hundreds of hours of on-location shooting across Italian villages, historic city centers, and forgotten countryside estates.

---

Dataset Creation

Curation Rationale

Architectural heritage photography is a niche but valuable domain for computer vision. Existing general-purpose datasets lack the granularity needed for architectural period estimation, material classification, style recognition, and heritage condition assessment. This dataset fills that gap by providing expert-annotated architectural photographs with detailed metadata suitable for fine-tuning vision-language models and training specialized classifiers.

Source Data

Initial Data Collection

Photographs were captured in situ across multiple countries and architectural periods. Images were selected for:

• Clear visibility of the primary architectural element
• Representative examples of specific styles and periods
• Variety of materials, conditions, and geographic origins
• High resolution suitable for detail analysis

Annotation Process

Metadata was generated using GPT-4o vision model with a specialized architectural history prompt. Each image was analyzed for:

1. Visual element identification (facade, door, window, arch, balcony, gate, etc.)
2. Architectural style classification based on visible features
3. Construction period estimation from stylistic cues and materials
4. Material and condition assessment
5. Keyword generation for broad discoverability

Annotations were validated against domain-specific classification dictionaries and cross-checked for consistency.

Personal and Sensitive Information

This dataset contains no personal information. All photographs depict buildings and architectural elements in public spaces. No people are identifiable in the images.

---

Considerations for Using the Data

Social Impact

This dataset supports:

• Digital preservation of architectural heritage
• Educational tools for architectural history and design
• Restoration planning through visual reference collections
• Cultural heritage documentation in endangered or changing urban environments

Discussion of Biases

• Geographic bias: Images are primarily from European and Mediterranean regions. Coverage of Asian, African, and Latin American architecture is limited.
• Period bias: Older, more visually distinctive periods (medieval, baroque, art_nouveau) may be overrepresented compared to generic modern architecture.
• Condition bias: The collection skews toward well-preserved or photogenic examples. Severely damaged or mundane contemporary buildings are underrepresented.
• Tourist bias: Many photos were taken at heritage sites popular with tourists, potentially overrepresenting monumental architecture.
• Element bias: Facades are overrepresented (30/50) compared to other architectural elements like windows (2) or balconies (1).

Other Known Limitations

• Construction period estimates are visual approximations based on architectural style and may not match exact documented dates.
• Location hints are inferred from architectural features and may be incorrect for imported styles or colonial architecture.
• The dataset does not include precise GPS coordinates or exact building addresses.
• Some modernist and contemporary buildings may be difficult to distinguish from similar generic styles.
• Color annotations are subjective and based on visual assessment rather than precise colorimetric measurement.

---

Changelog

Version	Date	Changes
1.0	2026-06-14	Initial release — 50 images with full metadata, compiled README, updated licenses

---

Keywords for Discovery

architecture, facades, doors, windows, walls, arches, gates, balconies, columns, historical buildings, French architecture, Italian architecture, European architecture, rustic, vintage, weathered wood, stone masonry, wrought iron, church doors, cathedral entrances, medieval, Gothic, Renaissance, Baroque, Neoclassical, Romanesque, Tuscan, Paris, Rome, Athens, texture, patina, architectural details, travel photography, generative AI training, computer vision dataset, 3D textures, interior design AI, heritage preservation, barn doors, iron gates, wooden doors, stone archways, building materials, restoration, preservation

---

Contributions

Contributions are welcome! If you have architectural photographs with verified metadata, please open a discussion or pull request on the Hugging Face Hub.

---

Last updated: June 14, 2026