FinWAIID - AI enhanced identification of fin whales using drone imageryi-MARSYS student research group wins Teaching Lab funding
11 July 2024, by Alexander Rychwalski, Kilian Huß & Helena Herr

Photo: Dr. Helena Herr, IMF UHH
Alexander Rychwalski and Kilian Huß have won a €10,000 grant for their student research group from the 'Digital and Data Literacy in Teaching Lab'.
Last Tuesday, iMARSYS students Alexander Rychwalski and Kilian Huß successfully presented and defended their project proposal FinWAIID at the latest UHH-ISA 'Digital and Data Literacy in Teaching Lab’ science slam.
The goal of FinWAIID is to develop a deep learning approach for automated photo identification using vertical aerial drone imagery of fin whales (Balaenoptera physalus).
Photo identification (photoID) is a vital tool for the study of whale populations. Traditional matching techniques are time-consuming and have become increasingly inefficient with growing databases. Advances in machine learning have demonstrated efficient performance of automated individual identification, but are restricted to lateral images of the dorsal fin. Increasing availability of drone imagery create a viable opportunity for vertical photoID based on the analysis of dorsal pigmentation patterns. So, Alex and Kilian will develop a deep learning-based framework for identifying individual fin whale in drone footage. They will aim to develop an aerial photoID approach by leveraging a semi-supervised workflow and employing a deep convolutional neural network architecture for (human) facial recognition for individual whale discrimination. This method is envisaged to aid rapid population assessment for whale conservation and management.
Alex and Kilian’s mentor, Dr. Helena Herr, really looks forward to the project results, as the scientific community already welcomed the approach last May, when it was presented to the Scientific Committee of the International Whaling Commission (IWC).