My AHRC-RLUK Skilled Observe Fellowship: 4 months on


In August 2022 I began work on a undertaking to research the legacies of curatorial voice within the descriptions of incunabula collections on the British Library and their future reuse. My analysis is funded by the collaborative AHRC-RLUK Skilled Observe Fellowship Scheme for tutorial and analysis libraries which launched in 2021. As a part of the first cohort of ten Fellows I embraced this chance to have interaction in practitioner analysis that advantages my establishment and the broader sector, and to advertise the function of library professionals as essential analysis companions.

The general goal of my Fellowship is to display new methods of working with digitised catalogues that will additionally enhance the discoverability and usefulness of the collections they describe. The main target of my analysis is the Catalogue of books printed within the fifteenth century now on the British Museum (or BMC) printed between 1908 and 2007 which describes over 12,700 volumes from the British Library incunabula assortment. Through the use of computational approaches and instruments with the information derived from {the catalogue} I’ll acquire new insights into and interpretations of this precious useful resource and allow its reuse in up to date on-line sources. 

Titlepage to volume 2 of the Catalogue of books printed in the fifteenth century now in the British Museum, part 2, Germany, Eltvil-Trier

BMC quantity 2 titlepage

This analysis thought was impressed by a latest collaboration with Dr James Baker, who can be my mentor for this Fellowship, and was additional developed in conversations with Dr Karen Limper-Herz, Lead Curator for Incunabula, Adrian Edwards, Head of Printed Heritage Collections, and Alan Danskin, Collections Metadata Requirements Supervisor, who assist my analysis on the Library.

My Fellowship runs till July 2023 with Fridays being my foremost analysis days. I started by learning the historical past of {the catalogue}, its association and the construction of the merchandise descriptions and their relationship with completely different on-line sources. Total, the principle focus of this primary section has been on producing the textual content knowledge required for the computational evaluation and investigations into curatorial and cataloguing apply. This work concerned new digitisation of {the catalogue} and quite a lot of experimentation utilizing the Transkribus AI-empowered platform that proved best-suited for bettering the structure and textual content recognition for the digitised photographs. Over the past two months I’ve massively benefited from the experience of my colleague Tom Derrick, as we labored collectively on creating the coaching knowledge and constructing construction fashions for the incunabula catalogue photographs.

An image from Transkribus Lite showing a page from the catalogue with separate regions drawn around columns 1 and 2, and the text baselines highlighted in purple

Structure recognition output for pages with solely two columns, together with textual content baselines, considered on Transkribus Lite


An image from Transkribus Lite showing a page from the catalogue alongside the text lines

Textual content recognition output after making use of the mannequin educated with annotations for two columns on the web page, considered on Transkribus Lite


An image from Transkribus Lite showing a page from the catalogue with separate regions drawn around 4 columns of text separated by a single text block

Structure recognition output for pages with combined structure of single textual content block and textual content in columns, considered on Transkribus Lite

While the information preparation section has taken longer than I had deliberate because of the assorted structure of {the catalogue}, this has been an essential a part of the method because the undertaking outcomes are depending on utilizing the very best quality textual content knowledge for the incunabula descriptions. The subsequent section of the analysis will contain the segmentation of the data and extraction of related info to make use of with a variety of computational instruments. I’ll report on the progress with this work and the subsequent steps early subsequent yr. Watch this area and do get in contact if you want to be taught extra about my analysis.

This blogpost is by Dr Rossitza Atanassova, Digital Curator for Digitisation, British Library. She is on Twitter @RossiAtanassova  and Mastodon @[email protected]


Leave a Reply

Your email address will not be published. Required fields are marked *