Home Damage, Three Frames (10,000 sets)

Home Damage, Three Frames (10,000 sets)

IMAGE

COMMERCIAL

LABELED

Each submission is three photographs of one defect in a contributor's home somewhere in Kled's global base: the whole room or wall, then closer, then the defect itself, with the issue type named (damp, crack, leak, peeling paint, broken tile). Close-up-only sets and unidentifiable issues do not get through review, and contributors are told not to create damage or include anything that identifies their address. Insurance and inspection models are trained on claims imagery that insurers do not license, and the open alternatives are small crack datasets; the wide-to-close structure here teaches a model to localize a defect in context rather than classify a crop.

Volume

10,000+

Photo sets

Contributors

10,000+

Regions

100+ countries

Updated

Technical specs

Technical specs

What’s in the collection and how it’s delivered.

Data modality

Images

Annotated

Labeled (defect type)

Exclusivity

Non-exclusive rights

Capture tool

Phone camera (in app)

Use cases

Property defect detection and localization, insurance claims triage, multi-scale visual grounding

Format

JPG / PNG / HEIC

Resolution

Native phone camera resolution (JPG, PNG, HEIC)

Total volume

10,000+ photo sets (approx. 90 GB)

Regions

100+ countries

Purity level

99%+ passed QA review

Ready to test
this dataset?

Ready to test
this dataset?

Get a sample set to evaluate, or talk to our data team about volume, exclusivity, and custom collection.

Get a sample set to evaluate, or talk to our data team about volume, exclusivity, and custom collection.