PowerPoint-Präsentation


1

Amir Moghaddass Esfehani
Campus Library

ADHO DH2019 Workshop 
"Towards Multilingualism In Digital Humanities: Achievements, 
Failures And Good Practices In DH Projects With Non-latin
Scripts"

No text – no mining. 
And what about dirty OCR? 



2

No text – no mining. 
And what about dirty OCR? 

• OCR
• Metadata
• Tools & APIs

No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani



3

Dirty OCR: Layout & Text

No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani

㊅
@
問
靈你’身栎物肩
“ 1 ， ？ \ ^ , 41安5 ~
10，
與I
神的
蓋自神你
也^重孤确触&何^)!
魂
何
耶
0
原
耶。
荷本狗' 0

耶。造等答白，
0我的'曰，小
愁的物眞子
答爲
承
是
魂
也。
地如
~萬此
^ # ^
|+之教
之神也。
人

Precision = 0,448
Recall = 0,371
F-measure = 0,272
Error rate = 58,3%



4

„Discovery“ Of A Chinese Ice Age

No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani

夾

襖綿

冰



5

„Discovery“ Of A Chinese Ice Age

No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani

夾 + 冰 + 綿 + 襖

1750

來 + 永 + 秀 +澳
| | | |



6

Metadata
嚴如熤 Yan Ruyi: 苗防備覽: [22卷] Miao fang bei lan [22 juan]. 紹義堂, Daoguang 23 
[China, 1843]. https://nbn-resolving.org/urn:nbn:de:bvb:12-bsb11123105-5

No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani

Descriptive
Structural

Rigths

Technical

OCR

MDZ > OPAC > DEAC > DDB > ZVDD > Europeana

MARC
RDF
IIIF

Retrieval
MARC
RDF
METS/
MODS
IIIF

cortex
EDM

EDMMETS/
MODS

MARC

non-latin script



7

Chinese Text Project: OCR 

No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani

Sturgeon, Donald (2018): Large-scale 
Optical Character Recognition of Pre-
modern Chinese Texts. International 
Journal of Buddhist Thought and 
Culture (2), p. 11-44.

https://digitalsinology.org/zh/wiki/Fil
e:Ctext-ocr.png



8

Chinese Text Project / MARKUS: Tools & APIs

No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani

Text reuse Regex NER (MARKUS)



9No text – no mining. And what about dirty OCR? ADHO 08.07.2019 Amir Moghaddass Esfehani

Thank you!

AMIR MOGHADDASS ESFEHANI
Campus Library
Freie Universität Berlin

amir.moghaddass@fu-berlin.de


	No text – no mining. �And what about dirty OCR? 
	No text – no mining. �And what about dirty OCR? 
	Dirty OCR: Layout & Text
	„Discovery“ Of A Chinese Ice Age
	„Discovery“ Of A Chinese Ice Age
	Metadata
	Chinese Text Project: OCR 
	Chinese Text Project / MARKUS: Tools & APIs
	Foliennummer 9