ID 2506: Video-based Re-Identification Dataset for Polar Bears and Other Species
Video-based animal re-identification has the potential to outperform image-based approaches by leveraging not only spatial but also temporal features. However, progress in this field is severely hindered by the lack of publicly available datasets. To address this, our group has developed one of the first datasets for wildlife re-identification based on video: PolarBearVidID [1]. For this Master’s project, we aim to extend this dataset, utilizing already recorded and identity-annotated video footage of additional polar bears and other species. The key challenge is to process and curate identity-annotated video sequences, making them available for further research in this area.
[1] Zuerl, Matthias, et al. “PolarBearVidID: A video-based re-identification benchmark dataset for polar bears.” Animals 13.5 (2023): 801.
Tasks:
- Creating identity-annotated video sequences for additional polar bears and other animals.
- Potential follow-up Master’s thesis with further tasks in re-identification research.
Requirements:
- Strong programming skills (Python, PyTorch).
- Experience with deep learning, particularly in computer vision (CNNs, action recognition, or pose estimation).
- Interest in wildlife conservation and ecological modelling.
- Ability to work independently and collaborate with an interdisciplinary research team.
Supervisors
Please use the application form to apply for the topic. We will then get in contact with you.