A Deep Learning Algorithm for Tracking Laboratory Animals in the “Morris Water Maze”

The “Morris water maze” presents a crucial challenge for tracking the moving small laboratory animal. Previously, we had solved the problem of obtaining the trajectory by analyzing video data using computer vision methods. To develop a more universal algorithm, this paper presents a trajectory construction algorithm based on the YOLO11 deep learning model, which is used for frame-by-frame object detection. The developed method shows promising results in overcoming the shortcomings of the previous approach.