Scientists use state-of-the-art artificial intelligence to extract complex information from vast collections of museum specimens.
A Cardiff University team is using state-of-the-art techniques to automatically segment and capture information from museum specimens and make significant data quality improvements without human intervention.
They have worked with museums across Europe, including the Natural History Museum in London, to refine and validate their new methods and contribute to the colossal task of digitizing hundreds of millions of specimens.
With more than 3 billion biological and geological specimens housed in natural history museums around the world, digitizing museum specimens, in which the physical information of a particular specimen is transformed into a digital format, has become a task of increasingly important for museums as they adapt. to an increasingly digital world.
A treasure trove of digital information is invaluable to scientists trying to model the past, present and future of organisms and our planet, and could be key to addressing some of the greatest societal challenges facing our world today’ today, from the conservation of biodiversity to the fight against climate change. finding new ways to deal with emerging diseases like COVID-19.
The scanning process also reduces the amount of manual handling of specimens, many of which are very delicate and susceptible to damage. Having appropriate data and images available online can reduce risk to the physical collection and protect specimens for future generations.
In a new article published today in the journal Machine vision and applicationsthe Cardiff University team has taken a step forward to make this process cheaper and faster.
“This new approach could transform our digitization workflows,” said Laurence Livermore, deputy digital program manager at the Natural History Museum in London.
The team created and tested a new method called image segmentation, which can easily and automatically locate and delineate different visual regions on images as diverse as microscope slides or herbarium sheets with a high degree of accuracy.
Automatic segmentation can be used to focus the capture of information from specific regions of a slide or sheet, such as one or more of the labels pasted on the slide. It can also help perform important quality control on images to ensure digital copies of specimens are as accurate as possible.
“In the past, our scanning was limited by the speed at which we could manually verify, extract and interpret data from our images. This new approach would allow us to augment some of the slower parts of our scanning workflows and to make crucial data more readily available to climate change and biodiversity researchers,” Livermore continued.
The method was trained and then tested on thousands of images of microscope slides and herbarium sheets from different natural history collections, demonstrating the adaptability and flexibility of the system.
Images contain key information about the microscope slide or herbarium sheet, such as the specimen itself, labels, barcodes, color charts, and institution names.
Typically, once an image has been captured, it then needs to be checked for quality control purposes and the tag information saved – a process currently done manually, which can be time and resource intensive.
Lead author of the new study, Professor Paul Rosin, from Cardiff University’s School of Computing and Computing, said: “Previous attempts to segment images of microscope slides and herbarium sheets were limited to images from a single collection.
“Our work leveraged the multiple partners of our large European project to create a dataset containing examples from multiple institutions and shows how well our AI methods can be trained to process images from a wide range of collections.
“We believe this method could help improve the workflows of staff working with natural history collections to dramatically speed up the digitization process at very little cost and resources.”
The microscope slides were provided by the Natural History Museum, Royal Botanic Gardens, Kew and Naturalis Biodiversity Center, while the herbarium sheets were provided by the National Museum Wales, Muséum National d’Histoire Naturelle, Museum für Naturkunde, Finnish Museum of Natural History, Meise Botanic Garden, Natural History Museum and Naturalis Biodiversity Center.