In this repository i've gathered multiple data sources for experiments on architecture and design using image processing such as GANs or CNNs and many other techniques Feel Free to send more DataSources (Makesure it's open sourced and link is not broken)
Data is like ingredients to machine learning models, algorithms are like the pan you're preparing it, and GPUs would be the fire! Pick up the right ones and make your own recipe
10113 images from 25 architectural styles (Multiple Sizes)
https://www.kaggle.com/dumitrux/architectural-styles-dataset/version/1
This dataset contains 12.3k+ images of architecture organized into 9 different categories based on time periods.
https://www.kaggle.com/gustavoachavez/architectural-styles-periods-dataset
78.5GB Total size of images from renthop.com NYC - Apartament pictures (Sample + Torrent Download - Multiple Sizes)
https://www.kaggle.com/c/two-sigma-connect-rental-listing-inquiries/data
506 Building Facades & corresponding Segmentations (256x256) PS: From the famous Pix2Pix Article
https://www.kaggle.com/balraj98/facades-dataset
5000 Images annotated into over 80 floorplan object categories
https://github.com/CubiCasa/CubiCasa5k
https://zenodo.org/record/2613548#.YAwn0-hKjIV
510 Images of Black & White Floorplans (Multiple Sizes)
https://github.com/gesstalt/ROBIN
122 Images of scanned floor plan documents divided in 4 different subsets regarding their origin and style
http://dag.cvc.uab.es/resources/floorplans/
1000 Images Total --> 10 Different floorplans with 100 different layouts each (Multiple Sizes)
http://mathieu.delalandre.free.fr/projects/sesyd/symbols/floorplans.html
103 Original Paired Images, and about 300 Paired Augmented Images (256x256) From: Enabling alternative architectures - Harvard GSD
https://www.nathanpeters.us/enabling-alternative-architectures
https://github.com/nonoesp/pix2pix-floorplans-dataset
80787 - Building boundry (black box) (256x256) From: Data-driven Interior Plan Generation for Residential Buildings (ACM Transactions on Graphics (SIGGRAPH Asia))
http://staff.ustc.edu.cn/~fuxm/projects/DeepLayout/index.html
1.6 GB --> Rental data (5.33 million all over Japan) 210 GB --> Image data (83 million files): floor plan image, room view, etc. 4.5 GB --> Property data of rentals and sales (5.33 million all over Japan)
https://www.nii.ac.jp/dsc/idr/en/lifull/
https://github.com/Littel-Laboratory/homes-dataset-tools
There are 35,126 2D floor plans with 252,550 rooms, with mean of 7.14 and median of 7.0 rooms per house. Together with the category label of each room
https://github.com/teaganli/houseexpo/
An end-to-end pipeline to render an RGB-D-inertial benchmark for large scale interior scene understanding and mapping. Contains 20M images
3D-FRONT contains 18,797 rooms diversely furnished by 3D objects 7,302 furniture objects all come with high-quality textures
https://tianchi.aliyun.com/specials/promotion/alibaba-3d-scene-dataset
948 Images of Chairs (Multiple Sizes)
https://www.kaggle.com/rik8088/chairstool
4147 images (256x256)
https://www.kaggle.com/aishahsofea/interior-design
4447 Images of Furniture - 900 Images each Class (Bed, Chair, Sofa, Swivelchair, Table) - (Multiple Sizes)
https://www.kaggle.com/akkithetechie/furniture-detector
3475 Images of city street view separated image data set (Paired Dataset 512x256)
https://www.kaggle.com/zhangyunsheng/cityscapes-data
12076 Images of a paired dataset (256x256)
https://www.kaggle.com/hyyyrwang/buildings-dataset
151 Aerial Images of Boston City to aid Machine Learning for Aerial Image Labeling (1500x1500)
https://www.kaggle.com/balraj98/massachusetts-buildings-dataset
Is a large-scale photo-realistic dataset containing 3.5K house designs (a) created by professional designers with a variety of ground truth 3D structure annotations
https://github.com/bertjiazheng/Structured3D
SUNCG is a large-scale dataset of synthetic 3D scenes with dense volumetric annotations.
https://sscnet.cs.princeton.edu/
The repository contains over 300M models with 220,000 classified into 3,135 classes arranged using WordNet hypernym-hyponym relationships. ShapeNet Parts subset contains 31,693 meshes categorised into 16 common object classes (i.e. table, chair, plane etc.). Each shapes ground truth contains 2-5 parts (with a total of 50 part classes).
There are 2586 houses in total, x-y-z dimensions
https://github.com/facebookresearch/droidlet
https://www.kaggle.com/raypratik/door-dataset-for-mrcnn
9565 images of Church elements
https://www.kaggle.com/adeshg/architectural-heritage
8 classe of: there is no event, fallen tree, fire, flooding, bad road, traffic jam, garbage, traffic accident
https://www.kaggle.com/tunhunhminh/demodata
https://paperswithcode.com/datasets
https://www.kaggle.com/datasets
https://datasetsearch.research.google.com/