Skip to content

Dataset · hyperbots-inc

Synthetic degraded document dataset

hyperbots-inc/synthetic-bad-scans

120 corporate document images with various degradation levels for OCR testing.

What's inside

The dataset contains 40 images across three degradation levels (light, moderate, heavy) simulating scanners, faxes, and phone cameras. Each image includes clean text, bounding box layouts, and degradation parameters. The data is presented in English in a small volume (less than 1K examples).

What you can do with it

  • Evaluating and improving OCR system accuracy on damaged scans
  • Testing document image alignment and enhancement algorithms
Tasks
Image → text
Languages
English
Size
up to 1K rows
Downloads this month
538
Likes
50

How to load it

from datasets import load_dataset

ds = load_dataset("hyperbots-inc/synthetic-bad-scans")