Specimen Data Refinery

The SDR is concerned with digitisation pipelines for digital access to natural history collections

The SDR integrate machine learning, Artificial Intelligence, and human approaches to extract, enhance, and annotate data from digital images and records at scale. Many collections-holding institutions still need to digitise the bulk of their collections. Digitisation takes time and resources. One of the major challenges in digitising massive collections is finding ways of ensuring high-quality collections data can be processed at pace.

We use new technological approaches, such as computer vision, data mining and machine learning, to rapidly enhance minimal natural history specimen records using images (e.g. of labels, specimens or registers) and unstructured text at scale. These approaches will be largely automated and may support record enhancement by experts as well as members of the public (crowdsourcing).

SDR is part of the Synthesys+ project, which is a project of the DISSCo ESFRI (https://www.dissco.eu/)