Dataset · biglam
Illustrated pages of the Britannica encyclopedia
biglam/britannica-illustrated-pages
Over 115k illustrated pages from historical editions of the Britannica encyclopedia (1768–1929).
What's inside
The dataset contains images, tabular data, and English text, covering 1 to 10 million records by metadata. The main part includes 115,293 scanned illustrated pages from editions 1–14 of the encyclopedia (1768–1929), selected by a classifier from 975,345 pages (1,160 volumes from the Internet Archive). The second configuration stores classifier scores, OCR word counts, and provenance metadata for all 975,345 pages; 82% of the illustrated pages contain 100+ OCR words with embedded diagrams and engravings.
What you can do with it
- Training page classifiers and historical engraving and diagram detectors
- Testing OCR and complex layout analysis systems for antique printed publications
- Tasks
- Image → text, Image classification
- Languages
- English
- Size
- 1M–10M rows
- Downloads this month
- 963
- Likes
- 51
How to load it
from datasets import load_dataset
ds = load_dataset("biglam/britannica-illustrated-pages")