Biodiversity research through artificial intelligence

Naturalis Biodiversity Center,, and COSMONiO have started a collaboration to support and improve biodiversity research through artificial intelligence. The collaboration will focus on using modern machine learning techniques, such as deep learning, to provide easy tools to identify and search through biodiversity data. The tools will be opened up both to researchers and to the larger public with the aim of accelerating biodiversity research and data collection.

As the amount of collected and accessible biodiversity data, such as natural history specimens and observation data, is growing, there a is need for better tools to help identify the contents of this data. More and more biodiversity data have multimedia associated with them as a result of the ubiquitous presence of smartphones and the increased efforts to digitize natural history collections. Naturalis Biodiversity Center has one of the largest natural history collections and is one of the frontrunners in digitizing its collection. Stichting Natuurbank Nederland, through the web platforms and, have one of the largest high quality databases of nature observations in the world. Recent advances in machine learning technology in the form of deep learning open the possibility to use these very large databases to create advanced image recognition tools for automatically identifying species of these media. Knowledge for creating these tools will be provided by COSMONiO which is a specialist company in machine learning.

The collaboration will focus on two main application areas: natural history collections and observations of the living nature. For the natural history collections web services will be developed to detect (isolate) and identify species in digitized specimens. These web services can be used by other applications to support a wide array of tasks, for example record creation, database validation, and content based image search. Besides an interest to increase scientific knowledge about developing new machine learning methodology for very large natural history datasets, there is very much a focus on making this methodology available in a standardized and reliable manner.

The other important application area is the identification of species from pictures taken in the wild. These pictures pose a unique of set of problems as a result of the complex appearance of the species and the background in the picture. Again the methodology will be made available through web services which can be used by other developers and the public. For example the website will be enriched with species identification technology. For this application area an important goal is to support citizen science initiatives concerned with biodiversity research that could benefit from automatic species identification. Such efforts are expected to be increasingly important given the rapid changes that occur in biodiversity worldwide due to amongst other climate change. Another goal is to further the knowledge of the public about the natural world by helping them to access and understand its large diversity.

For more information the project, please contact Laurens Hogeweg (