Program

Segment 1 — Introduction [8:30—8:45 am]

Animal ReID: Challenges, Opportunities, and Impact

• About animal re-identification. Importance of ecological research, conservation, and wildlife management.

• Discuss real-world applications, e.g., biodiversity monitoring, animal behavior…

• Highlight the role of citizen science platforms like WildMe, iNaturalist, and GBIF.

• Key challenges, e.g., non-standardized evaluation, dataset scarcity, and ethical considerations.

• Interdisciplinary nature of the research, i.e., collaboration of CV, ML, and ecology communities.

Segment 2 — Methods and Tools [8:45—10:00 am]

From Classics to Cutting-Edge: Methods for Animal ReID [20m + 5m Q&A]

• Evolution of animal ReID methods, from traditional computer vision approaches through modern deep learning techniques to their combination.

• Comparison of methods based on specific species, use cases, and data availability, explaining why classical methods are still viable for some scenarios.

Building Bridges: Unifying Wildlife Datasets for ReID [15m + 5m Q&A]

• Need for unifying and standardizing wildlife datasets

• Complexities around sharing data for sensitive species trade off with the benefit of multi-species training, and potential paths forward that can support model development without putting species at risk (i.e., Wildme and private data collection vs Open-source data collection)

• Existing efforts and strategies for creating a global open-access repository of annotated data.

• Existing tools and frameworks for handling diverse datasets.

Past, Present, and the Future Challenges  [10m + 5m Q&A]

• Open-set vs. closed-set recognition and domain adaptation challenges, including detecting unknown individuals and adapting to habitat, sensor, and lighting shifts.

• Scalability, ethics, and collaboration. lightweight on-device inference, streaming vs. batch monitoring, balancing open data with species protection, and fostering community benchmarks and contributions

Coffee Break [10:00—10:45 am]

Segment 3 — Case Studies [10:45—11:15 am] 3*7m+3m Q&A