Skip to Main content Skip to Navigation
Journal articles

La correction participative de l’OCR par crowdsourcing au profit des bibliothèques numériques

Abstract : For their digitization projects, libraries produce often OCR with errors which can be corrected by providers employing low cost labor. But libraries May also appeal to web volunteers (explicit crowdsourcing) or to a paid crowd (like Amazon Mechanical Turk marketplace) or to users correcting OCR by playing games (gamification) or to internet users who don’t know that they are correcting OCR (implicit crowdsourcing like reCAPTCHA). Profitability of these experiments is compared.
Document type :
Journal articles
Complete list of metadata

https://hal.inrae.fr/hal-02631051
Contributor : Migration Prodinra <>
Submitted on : Wednesday, May 27, 2020 - 7:07:19 AM
Last modification on : Thursday, April 29, 2021 - 3:35:11 AM

Identifiers

  • HAL Id : hal-02631051, version 1
  • PRODINRA : 307678

Citation

Mathieu Andro, Imad Saleh. La correction participative de l’OCR par crowdsourcing au profit des bibliothèques numériques. Bulletin des bibliothèques de France, Ecole Nationale Supérieure des Sciences de l'Information et des Bibliothèques (ENSSIB), 2015. ⟨hal-02631051⟩

Share

Metrics

Record views

56