While a lot of progress have been made in individual identification in humans, individual identification of animals is less well developed, despite being of large relevance within scientific research and biodiversity conservation. Some species of animals have visible marks or individual patterns, but other species are visually “identical” to humans, which poses a serious challenge for individual identification. In this project, a large dataset of individually marked birds are used as a basis of training machine learning models for individual identification. The goal is to build a fully automated system in the field where birds are identified in real time using stationary and/or moving video cameras. The data collection takes place in a world unique ecological research facility on the island of Stora Karlsö, the Karlsö Auk Lab, http://www.balticseabird.com/auklab/