HTR in the making : En studie av hur Handwritten Text Recognition görs vid tre svenska arkivverksamheter

Detta är en Master-uppsats från Lunds universitet/Avdelningen för ABM, digitala kulturer samt förlags- och bokmarknadskunskap

Sammanfattning: In the current archival science paradigm, archives implement AI-technologies to provide content as structured, machine-readable data. One of these technologies is Handwritten Text recognition (HTR) which can transcribe handwritten text. Thus, HTR turn raw digitized archival documents into machine-readable format. Technology is not developed in a vacuum, but is shaped by heterogeneous factors, therefore it is relevant to shed light on how HTR technology is implemented in the Swedish archival sector. The purpose of this thesis is therefore to create an understanding of the relationship between Handwritten Text Recognition and archives by studying projects where this AI technology is being applied. By Actor-Network Theory the analysis is looking for agencies, mediators, and translation within the shaping processes of the HTR technology. The material span three projects carried out by Swedish archives managements, at Uppsala University Library, the Swedish National Archives and the Swedish National Heritage Board. Data was collected by interviewing archivists involved in the projects and from qualitative document analysis of the projects’ descriptions. The analysis provides an indication that HTR is particularly valued for future data driven research and for the future archival users for whom HTR is seen as a prerequisite. Making HTR through citizen science and crowdsourcing is also understood as a central factor for how HTR is being made within the projects. Continually the results show that HTR is made nationally within Sweden and Finland, partly by the exchange of experiences and by the HTR infrastructure Transkribus. HTR is understood, not primarily as a tool to produce results, but as a methodological updating within the institutions, and a step towards future expectations. The conclusions can therefore in many ways be related to the archive-as-data paradigm where the archives studied take part in the ongoing discourse of change and with making themselves relevant for the future.

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)