Dataset · ZhihaoNan
Synthetic web UI screenshots with annotations
ZhihaoNan/AtomBlock-WebUI
Around 9,700 full-size webpage screenshots with UI element annotations in YOLO format for interface detection.
What's inside
Contains ~9,700 synthetic English webpage screenshots and bounding box annotations in YOLO format. The layout covers 14 element categories, ranging from basic components like buttons and input fields to major semantic blocks (navigation, sidebar, footer). Data generated via LLM-enhanced HTML rendering and headless browser screen capture.
What you can do with it
- Training object detectors (such as YOLO family models) for recognizing web interface elements
- Building UI agents and automated web page structure parsing systems
- Tasks
- Object detection
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 530
- Likes
- 47
How to load it
from datasets import load_dataset
ds = load_dataset("ZhihaoNan/AtomBlock-WebUI")