Search papers, labs, and topics across Lattice.
University of Bristol
5
0
5
Achieving a remarkable 99.40% accuracy on a highly imbalanced dataset, this research uncovers critical gaps in long-tail recognition performance under domain shifts.
Simple aggregation of predictions outperforms complex models in detecting leprosy in wild chimpanzees, highlighting the importance of data handling in wildlife health monitoring.
The PanAf-SBR dataset reveals that fine-grained social behaviour recognition in wild great apes can be significantly improved through targeted cross-dataset pre-training.
Automated identification of individual animals can only be effective if it aligns with ecological questions and data practices, not just algorithmic accuracy.
Cows in a crowd? This new pipeline uses SAM and contrastive learning to achieve 94.82% re-identification accuracy, blowing away previous methods.
