Say cheese, Cheeta!

Great ape watching station at Zoo Leipzig uses face recognition technology by Fraunhofer

Press Release /

On March 23, 2016, Zoo Leipzig will be presenting its new interactive great ape watching station to the public. Part of the station will be the intelligent face recognition software developed by Fraunhofer, which will be used for automatic identification of individual chimpanzees living in Pongoland (what the ape enclosure at Zoo Leipzig is called). Visitors at Zoo Leipzig may use the system to request information about each individual animal in the compound.

© Zoo Leipzig
Thanks to Fraunhofer’s face recognition software, visitors at Zoo Leipzig may request information about each member of the chimpanzee group.


Natascha is 35 years old. Her distinctive visual features are these little bright spots on both of her lips, and a white chin beard. Natascha is one of the apes living in Pongoland which can now be electronically captured by means of the new system, giving visitors the opportunity to learn that Natascha is very close friends with Dorien, that she is the mother of Frodo, and that she has been at Zoo Leipzig since 2001.

And this is how the new system works: By means of simple directional navigation, a swivel-mounted camera can be directed towards individual animals in the compound. On a screen located in the watching station, each of the animals captured is then marked by a frame around its face, and the name of the chimpanzee is displayed. Visitors may then request information about each animal selected.


In the project, Fraunhofer IDMT was responsible for providing the face recognition technology. To this purpose, the Fraunhofer experts fed a video recognition software with the facial characteristics of individual apes, allowing to identify each animal by means of distinctive facial features, such as individual patterns of wrinkles under the chimpanzee’s eyes, or specific colorings.

Up to a certain angle of view, the face recognition technology by Fraunhofer is tolerant against the position of the ape’s face. To identify an ape, the system must be able to recognize both of the ape’s eyes, as they function as position markers. For each of the 17 chimpanzees at Zoo Leipzig to be unambiguously identifiable, the system had to be fed with 20 to 30 pictures of each animal (which varied in terms of perspective, light, etc.). If the group is joined by a new member, the system can easily be fed with new data.

During the project, the researchers were confronted with a number of challenges. "Unlike human beings, apes do not always cooperate when taking part in a photo shoot. They do not sit still when they are supposed to, and they rarely look straight into the camera", says Alexander Loos, project manager at Fraunhofer IDMT. "Another challenge was to ensure that the face recognition works reliably under different ambient lighting conditions, and also when the chimpanzees make sudden, unexpected moves. Chimpanzees are very playful animals, so we had to make sure they can be recognized by the system even if their faces are partially occluded by toys, ropes, sticks, or other objects."


The ape watching station is an outcome of a collaboration of Zoo Leipzig, Max Planck Institute for Evolutionary Anthropology, Fraunhofer IIS, and Fraunhofer IDMT. The one-year R&D project was funded by Klaus Tschira Stiftung.

The scientific basis of the project was laid in an earlier R&D project named SAISBECOSemi-Automated Audiovisual Species and Individual Identification System for Behavioral Ecological Research and Conservation«), featuring participation of Max Planck Institute for Evolutionary Anthropology, the University of Bristol, Fraunhofer IIS, and Fraunhofer IDMT. In this project, the researchers developed technologies for identifying great apes in the wild through automatic evaluation of video and audio data recorded.


Last modified: