Hugging Face's logo

Datasets:
tachiwin
/
ocr-test-challenging-rescanned

Tasks:
Image-to-Text
Modalities:
Tabular
Text
Formats:
parquet
Size:
< 1K
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
Dataset Viewer
Auto-converted to Parquet Duplicate
page_id
stringlengths
70
70
pdf_hash
stringclasses
17 values
page_number
int32
0
350
text
stringlengths
59
3.28k
uncommon_char_score
float32
0.5
1
image
dict
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800_p0082
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800
82
"_Los pronombres personales con el enclítico_ _**=ñe**_ \n\n68 \n\nun cambio de tono en la sílaba(...TRUNCATED)
0.548153
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOz9a3fjSrKmCZqZA6QUsTN3nlOnK0/Xqj(...TRUNCATED)
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800_p0070
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800
70
"56 \n\n_Los pronombres personales dependientes_ \n\nb. xtyoꞌ(m) ca.ntyoꞌ(b.m) _cuello de cuello(...TRUNCATED)
0.574046
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOz9XXPkOs7niwKglGnXWv289OzZs+NcTZ(...TRUNCATED)
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800_p0054
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800
54
"40 \n\n_Los pronombres personales dependientes_ \n\n(21) ma-cwaꞌ(m.m) _come_ ma-cwa=ꞌ(m.a-b) _c(...TRUNCATED)
0.516349
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOz97ZbcOI6uDQMgFZFpV/f09Mzea6/17P(...TRUNCATED)
6fe4c85920beb3431e0032dcf1d9093be2c7fa6461fa2dfc474604f255558e35_p0012
6fe4c85920beb3431e0032dcf1d9093be2c7fa6461fa2dfc474604f255558e35
12
"kjimachjénre tki̱. \n\nTsubáre nga nikitjusúnre xukúcha tjitsú xu̱ ta̱ nga ña'án xki̱, (...TRUNCATED)
0.589678
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOy9bXfjOM6uC4CSnVTXzJ5Zez/7wzn//w(...TRUNCATED)
6fe4c85920beb3431e0032dcf1d9093be2c7fa6461fa2dfc474604f255558e35_p0013
6fe4c85920beb3431e0032dcf1d9093be2c7fa6461fa2dfc474604f255558e35
13
"- ac) Tuxi manga fitjusún kjuakixí'e tjún xu̱ ta̱ xi ni'ín inchamáre ña'án xki̱, ngat'a (...TRUNCATED)
0.550054
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOz9bXPjSJImart7BEhldXXPzK7t2WO2Zs(...TRUNCATED)
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800_p0188
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800
188
"174 \n\n_Las formas singulares y plurales de los verbos_ \n\n(34) cwi-co (m.m) _se quema, se queman(...TRUNCATED)
0.585255
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOz965rbyK61CwIIUlLaVev09WF39/1f4N(...TRUNCATED)
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800_p0040
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800
40
"26 \n\n_Los pronombres personales independientes_ \n\ntercera persona del plural y el clítico **=y(...TRUNCATED)
0.515853
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOzda3vjuJIt6BUBUnJW7e6ZM898mP//A+(...TRUNCATED)
3bdde4164c4c074a93373a6e7dc8f91163f862a694a88be2cffc77d47e291117_p0007
3bdde4164c4c074a93373a6e7dc8f91163f862a694a88be2cffc77d47e291117
7
"A es xmod ñal ban rë men 7 Ziyit. Ley reꞌ ngurulë “Leyes de Reforma”, no nzhudiꞌzhe: \n\(...TRUNCATED)
0.596532
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOy94XrbyI6tvYCinD17zvnu/z5nzu62xQ(...TRUNCATED)
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800_p0346
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800
346
"332 \n\n_Textos_ \n\n_―Muchas gracias, amigo ―le dijo el ratón―. Voy a sembrar mañana y pas(...TRUNCATED)
0.558345
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOz97XLcSM+uCwKZZEl29/vsWWtmRUzM+R(...TRUNCATED)
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800_p0293
3d6086b073801b95457ebabb27f9f2ba5e0f38b235a40da5a6e7f60de323a800
293
"279 \n\n_Un verbo o frase verbal_ \n\n(4) _**Enfoque (sujeto) Predicado-Sujeto**_ a. Calueꞌñee(...TRUNCATED)
0.502716
{"bytes":"iVBORw0KGgoAAAANSUhEUgAABqQAAAvMCAIAAACYwgGYAAEAAElEQVR4nOz925bjuq6miwKgZEdmH7Nqrou126qHrl(...TRUNCATED)
End of preview. Expand in Data Studio

ocr-test-challenging-rescanned

Re-scanned pages from the original tachiwin/ocr-test-challenging-3 dataset. 200 pages were printed on physical paper and scanned back to digital images, providing a real-world OCR degradation benchmark.

Columns are identical to the original dataset:

  • page_id — unique page identifier
  • pdf_hash — hash of the original PDF document
  • page_number — page number within the original document
  • text — original ground-truth text
  • uncommon_char_score — score indicating character rarity
  • image — the scanned page image

Usage 

from datasets import load_dataset
ds = load_dataset('tachiwin/ocr-test-challenging-rescanned', split='train')
Downloads last month
97
Number of rows:
200
Total file size:
727 MB