Datasets:

---

dresserman

/

kanops-open-retail-imagery

like 0

What’s in v0

• ~10.8k retail-scene images across 2014 / FullStores / Halloween2024
• All faces blurred; trademarks belong to respective owners.
• Metadata: MANIFEST.csv, metadata.csv, checksums.sha256
• License: Evaluation-only (see LICENSE)

Known limitations
• No CVAT labels yet (coming in v1)
• Some EXIF warnings are expected; content is stamped

How to use
• imagefolder: load images via datasets.load_dataset("imagefolder", data_dir="hf://datasets/dresserman/kanops-open-access-imagery/train")
• Join access: request via the Gated form on the main page and add basic information about yourself.

Changelog
• v0: initial release

Total Dataset is over 1m images, with 280,000 structured by specific seasonal event sub category (IE Pumpkins).