DAMM for the detection and tracking of multiple animals within complex social and environmental settings
G. Kaul, J. McDevitt, J. Johnson, A. Eban-Rothschild
The paper introduces DAMM, a pretrained object detector for localizing mice in complex environments, with robust performance in zero-shot and few-shot inference, and integrates it with SORT for tracking.
bioRxiv · Jan 18, 2024Animal Behavior & Cognition