Semi-Automated Audiovisual Species and Individual Identification System for Behavioral Ecological Research and Conservation
Great apes are on the brink of extinction. An essential part of effective biodiversity conservation management is population monitoring using acoustic and remote camera devices. This is not only very time-consuming, but a process prone to errors. Therefore, a new type of identification software was developed within the SAISBECO project. It is capable of searching through video and audio footage for sequences in which the apes are visible, identifying specific individuals, and sorting the images accordingly. The developed software detects in single images or videos which images show the faces of great apes. Whether gorilla or chimpanzee – the algorithms allow the identification and classification of the ape species. After that, according to typical facial characteristics, structures and outlines, the individuals of a group are identified.
The goal of this project was:
- Collection and annotation of a large set of video and audio recordings from the African rainforest.
- Development of tools allowing the automatic identification of species and individuals from audio and video data.
- Development of ecological models for biodiversity, demographic and population monitoring.
The final goal was the development of a semi-automated monitoring tool using passive audio and video recordings.