Classification of crowdsourced text correction

Autor:	Haimonti Dutta, Megha Gupta, Brian Geiger
Rok vydání:	2015
Předmět:	Information retrieval business.industry Computer science media_common.quotation_subject Optical character recognition computer.software_genre Column (database) Newspaper Text mining Font ComputingMethodologies_DOCUMENTANDTEXTPROCESSING Quality (business) business computer media_common
Zdroj:	CODS
DOI:	10.1145/2732587.2732619
Popis:	Optical Character Recognition (OCR) is a commonly used technique for digitizing printed material enabling them to be displayed online, searched and used in text mining applications. The text generated from OCR devices is often garbled due to variations in quality of the input paper, size and style of the font and column layout. This adversely affects retrieval effectiveness and hence techniques for cleaning the garbled text need to be improvised. This prototype system is expected to be deployed on historical newspaper archives that make extensive use of user text corrections.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::44baa129098fee7fce55d2cd7967c8d0 https://doi.org/10.1145/2732587.2732619 Zobrazit plný text záznamu