Search

HTR (Handwritten Text Recognition) technologies have progressed enough to offer high-accuracy results in recognising handwritten documents, even on a synchronous level. Despite the state-of-the-art algorithms and software, historical documents (especially those written in Greek) remain a real-world challenge for researchers. A large number of unedited or under-edited works of Greek Literature (ancient or Byzantine, especially the latter) exist to this day due to the complexity of producing critical editions. To critically edit a literary text, scholars need to pinpoint text variations on several manuscripts, which requires fully (or at least partially) transcribed manuscripts. For a large manuscript tradition (i.e., a large number of manuscripts transmitting the same work), such a process can be a painstaking and time-consuming project. To that end, HTR algorithms that train AI models can significantly assist, even when not resulting in entirely accurate transcriptions. Deep learning models, though, require a quantum of data to be effective. This, in turn, intensifies the same problem: big (transcribed) data require heavy loads of manual transcriptions as training sets. In the absence of such transcriptions, this study experiments with training sets of various sizes to determine the minimum amount of manual transcription needed to produce usable results. HTR models are trained through the Transkribus platform on manuscripts from multiple works of a single Byzantine author, John Chrysostom. By gradually reducing the number of manually transcribed texts and by training mixed models from multiple manuscripts, economic transcriptions of large bodies of manuscripts (in the hundreds) can be achieved. Results of these experiments show that if the right combination of manuscripts is selected, and with the transfer-learning tools provided by Transkribus, the required training sets can be reduced by up to 80%. Certain peculiarities of Greek manuscripts, which lead to easy automated cleaning of resulting transcriptions, could further improve these results. The ultimate goal of these experiments is to produce a transcription with the minimum required accuracy (and therefore the minimum manual input) for text clustering. If we can accurately assess HTR learning and outcomes, we may find that less data could be enough. This case study proposes a solution for researching/editing authors and works that were popular enough to survive in hundreds (if not thousands) of manuscripts and are, therefore, unfeasible to be evaluated by humans.

Many libraries offer free access to digitised historical newspapers via user interfaces. After an initial period of search and filter options as the only features, the availability of more advanced tools and the desire for more options among users has ushered in a period of interface development. However, this raises a number of open questions and challenges. For example, how can we provide interfaces for different user groups? What tools should be available on interfaces and how can we avoid too much complexity? What tools are helpful and how can we improve usability? This paper will not provide definite answers to these questions, but it gives an insight into the difficulties, challenges and risks of using interfaces to investigate historical newspapers. More importantly, it provides ideas and recommendations for the improvement of user interfaces and digital tools.

While online crowdsourced text transcription projects have proliferated in the last decade, there is a need within the broader field to understand differences in project outcomes as they relate to task design, as well as to experiment with different models of online crowdsourced transcription that have not yet been explored. The experiment discussed in this paper involves the evaluation of newly-built tools on the Zooniverse.org crowdsourcing platform, attempting to answer the research question: "Does the current Zooniverse methodology of multiple independent transcribers and aggregation of results render higher-quality outcomes than allowing volunteers to see previous transcriptions and/or markings by other users? How does each methodology impact the quality and depth of analysis and participation?" To answer these questions, the Zooniverse team ran an A/B experiment on the project Anti-Slavery Manuscripts at the Boston Public Library. This paper will share results of this study, and also describe the process of designing the experiment and the metrics used to evaluate each transcription method. These include the comparison of aggregate transcription results with ground truth data; evaluation of annotation methods; the time it took for volunteers to complete transcribing each dataset; and the level of engagement with other project elements such as posting on the message board or reading supporting documentation. Particular focus will be given to the (at times) competing goals of data quality, efficiency, volunteer engagement, and user retention, all of which are of high importance for projects that focus on data from galleries, libraries, archives and museums. Ultimately, this paper aims to provide a model for impactful, intentional design and study of online crowdsourcing transcription methods, as well as shed light on the associations between project design, methodology and outcomes.

The creation of the Artist Libraries Project was sparked by the observation that artist libraries are still not well known, yet many art historians are interested in this archive for the value it adds to understanding the person behind the artist and his or her creative process. The problem is that these libraries are rarely physically preserved. To remedy this dispersion, we built an online database and a website www.lesbibliothequesdartistes.org that house this valuable source in the form of lists of books and their electronic versions. First data on Monet's library have been made available, and several additional artist libraries from the 19 th and 20 th centuries are on the way for 2019. By gathering all these bibliographical data in a central database, it's possible to explore one library and to compare several. This article explains how we built the database and the website and how the implementation of those IT tools has raised questions about the use of this resource as an archive on the one hand, as well as its value for art history on the other.