Linking natural history collections

We describe tooling and infrastructure that enables direct and collaborative semantic annotation of field book content, without the requirement for full transcription. These tools enable a more streamlined approach to the creation of rich, integrated archives that can be interlinked with other cultural history resources in the field. In our use case we are annotating and enriching data from expeditions undertaken by the Committee for Natural History (1820-1850). The collection contains approximately 10,000 specimens and 8000 handwritten field book pages. Due its vast size and heterogeneity, a full text transcription is no viable option. Despite digitization, the collection has thus remained inaccessible to scholars and the general public. At the core of our paper is the Semantic Fieldbook Annotator (SFB-A) which enables researchers, collection holders and possibly also citizen scientist to interact with digitized natural history collections. The SFB-A allows users to draw bounding boxes, or Regions of Interest (ROIs), over the image scans to which semantic annotations, e.g. semantic classes of words, can be attached. Instead of transcribing all text, we use the SFB-A to annotate salient named entities in field notes of which the semantics are defined in a formal ontology. Examples are taxonomical names or geographical locations. By doing so we aim to save time and cost, and also preserve a direct link to the original document image.