KIWI@SIWI is a project that deals with the recognition of European bison (alias Wisent) by AI technology. The aim is to use the technology to better recognize the animals and thus support herd management. The project is carried out by the University of Siegen and NanoGiants GmbH. It has received funding from the NEXT.IN.NRW funding line.

Project Description

For sustainable management of free-ranging wisents, it is crucial to be able to uniquely identify individual animals. Since artificial markings such as ear tags are avoided, the project relies on natural, individual differences in the appearance of the animals.

This is where Artificial intelligence (AI) comes into play: a neural network is to automatically recognize characteristic features - especially in the facial area, but also other individual characteristics - based on photos and thus enable reliable differentiation of the animals.

The work is divided into several steps: First, a comprehensive image dataset of the individual European bison is curated, which serves as the basis for training the neural network. Building on this, a mobile application is to be developed with which users in the field can photograph animals in real time and identify them immediately. Testing and scientific evaluation will take place with the "Wisent Wildnis" herd.

Beyond the direct application, the project investigates what new possibilities arise for herd management, and to what extent image data is necessary to develop a robust and practical AI solution. The project thus also serves as a feasibility study for the use of similar methods in other animal species.