AI recognizes chimpanzees faces in the wild!

New collaboration between primatologists from our lab and experts in computer vision (from the Visual Geometry Group) results in a new method for identifying wild chimpanzees from raw videos. This manuscript is part of Daniel Schofield's DPhil project, and is fundamental for future work, such as automated extraction of behavioural features, as you can see here from the social network example provided in the open access paper published in Science Advances.



The method was developed by applying a convolutional neural network with a deep architecture to 14-year digital video archive from Bossou. Schofield and colleagues hope this method can be extenteded easily to other species and hopefuly will contribute directly to the conservation of endangered primates in the wild.