Texterkennung in Handschriften nicht-westlicher Sprachen

Baptiste Queuche, Mitbegründer von CALFA, einer KI- und NLP Lösung für orientalische Sprachen, stellt das Werkzeug in Online Lecture für die Mittelalterforschung vor.

© IMAFO, Historische Identitätsforschung

CALFA ist ein neues Texterkennungs-Werkzeug insbesondere für orientalische Sprachen. Es wird von einem Forschungsteam in Paris entwickelt und wurde bereits zur Verarbeitung von Dokumenten in Altgriechisch, Äthiopisch, Hebräisch und Syrisch eingesetzt. CALFA bietet zudem große generische Erkennungsmodelle an, die ohne Training für Arabisch und Armenisch verwendet werden können.

Baptiste Queuche aus dem CALFA-Team stellt dieses Texterkennungs-Tool auf Einladung des Instituts für Mittelalterforschung der ÖAW in einer Online Lecture on DIGITAL HUMANITIES vor. In "Text recognition for non-western languages in manuscripts: presentation of Calfa" haben die Teilnehmenden Gelegenheit, die maßgeschneiderten Modelle von CALFA kennenzulernen und herauszufinden, inwieweit das Tools für ihre Forschungsarbeiten nützlich sein könnte.

Weitere Informationen

Information

Termin:
28. Mai 2025, 15:00 bis 16:30 Uhr

Online

Kontakt für Fragen, Anregungen, Feedback:
Dr. Jan Odstrčilík
Institut für Mittelalterforschung der ÖAW

Name	Purpose	Storage duration	Type	Provider
CookieConsent	Remembers your consent to the use of cookies.	1 year	HTML	Web Consent
fe_typo_user	Assigns your browser to a session on the server. This only affects the content you see and is not evaluated or processed by us	-	HTTP	Web User

Name	Purpose	Storage duration	Type	Provider
_pk_id	Used to store a few details about the user like unique visitor ID.	13 months	HTML	Matomo-id
_pk_ref	Used to store information about the user's referring website.	6 months	HTML	Matomo-ref
_pk_ses	Short-term cookie to save temporary data from the visit.	30 minutes	HTML	Matomo-ses
_pk_cvar	Short-term cookie to save temporary data from the visit.	30 minutes	HTML	Matomo-cvar
_pk_hsr	Short lived cookie used to temporarily store data for the visit.	30 minutes	HTML	Matomo

Name	Purpose	Storage duration	Type	Provider
YouTube	A connection to YouTube will be established to view videos.	-	Connection	YouTube
SoundCloud	A connection to SoundCloud will be established to play audio files.	-	Connection	SoundCloud
Twitter	A connection to Twitter will be established to display tweets.	-	missing translation: type.	Twitter
_cs_c	Indicates whether the user has consented to ContentSquare tracking.	394 days	missing translation: type.	Spotify (ContentSquare)
_cs_id	Stores a unique user ID for ContentSquare session analysis.	394 days	missing translation: type.	Spotify (ContentSquare)
_ga	Used to distinguish users.	400 days	missing translation: type.	Google Analytics
_ga_BMC5VGR8YS	Used by Google Analytics to persist session state.	400 days	missing translation: type.	Google Analytics
_ga_S0T2DJJFZM	Used by Google Analytics to persist session state.	399 days	missing translation: type.	Google Analytics
_ga_ZWG1NSHWD8	Used by Google Analytics to persist session state.	400 days	missing translation: type.	Google Analytics
_ga_ZWRF3NLZJZ	Used by Google Analytics to persist session state.	400 days	missing translation: type.	Google Analytics
_gid	Used to distinguish users.	1 days	missing translation: type.	Google Analytics
_ScCbts	Stores temporary session or playback preferences.	6 days	missing translation: type.	Spotify
_scid	Spotify advertising ID used for analytics and retargeting.	395 days	missing translation: type.	Spotify
_scid_r	Spotify advertising ID used for analytics and retargeting.	395 days	missing translation: type.	Spotify
eupubconsent-v2	Stores the IAB Transparency & Consent Framework string.	364 days	missing translation: type.	IAB / Spotify
OptanonAlertBoxClosed	Saves the state of your data protection consent.	364 days	missing translation: type.	OneTrust
OptanonConsent	Saves the state of your data protection consent.	365 days	missing translation: type.	OneTrust
sp_adid	Spotify advertising identifier.	365 days	missing translation: type.	Spotify
sp_landing	Tracks which page the user landed on within Spotify.	1 days	missing translation: type.	Spotify
sp_m	Stores the user’s market region (Spotify).	399 days	missing translation: type.	Spotify
sp_t	Session token used for Spotify playback and access.	365 days	missing translation: type.	Spotify