==============

2021-05-11 Updated After a few days of system outage, Transkribus is back! We have rescheduled the workshop as follows. We sincerely look forward to your participation.

 

Date: May 29th (Sat), 11:00-12:30

The Zoom link that was previously distributed is no longer valid. To participate in this workshop, please fill out the registration form. [https://us02web.zoom.us/meeting/register/tZElcOugqzoqEtXw8WRJgfvz2hrVkgZSk9Os]

===============

Over the last few years, digital technology has allowed scholars to collate and research huge quantities of source materials very quickly, in particular using big data techniques. One of the simplest but most useful of these has been the plethora of Optical Character Recognition (OCR) software that has enabled researchers to search for various keywords in multi-volume historical sources and gain results almost instantly. However, these programmes have been almost exclusively focused on printed, published texts, leaving manuscripts and other handwritten materials left out. Yet recently, a new programme has been developed, Transkribus, which enables scholars to digitally search across handwritten source material in the same way as for printed texts with OCR. In this talk, I will introduce Transkribus and its main features, and explore how it can be used to reconstruct networks and connectivity in Islamic society, primarily using the example of documents found in archives in Malta.

Date&Time: May 29th, 11:00–12:30

Presentation: Alex Mallett

Discussion

Language: English

Venue: Zoom, Open to the public & Free, Pre-registration is required

Pre-registration: Please fill in the pre-registration form.

Co-organizer: Grant-in-Aid for Transformative Research Areas (A), “Connectivity Analyses by Digital Humanities Method” (Principal Investigator: Wakako Kumakura (ILCAA); 20H05830) & Grant-in-Aid for Transformative Research Areas (A), “The Ideas of the Muslim Community and State Systems” (Principal Investigator: Nobuaki Kondo (ILCAA); 20H05827)

活動記録

5月29日開催のワークショップにおいては、報告者のAlex Mallett氏が、手稿本の文字認識エンジン（HTR, Handwritten Text Recognition）であるTranskribusに関する機能と使い方について、マルタ島の文書史料の事例を用いて紹介した。質疑応答ではソフトの使い方に関する議論を中心に行われた。Transkribusは欧米諸語のモデルは多数公開されているものの、アラビア語のモデルは現時点において公開されていないため、Transkribusでアラビア語の文書を読ませるためには、教師モデルをつくるところから始めなくてはならない。他方、適合したモデルがすでにあれば、高い精度で文字認識が可能であることも共有された。