Influence of hyperparameters on the performance of deep learning-based microrobotic localization under phantom tissue
Johanna Hoppe, Richard Nauber, David Castellanos Robles, Jürgen W. Czarske, Mariana Medina‐Sánchez
This study evaluates how hyperparameters and model versions affect deep learning-based tracking of microrobots in phantom tissue, finding that YOLOv9 outperforms YOLOv8 and adjusting loss functions improves performance.
bioRxiv · Aug 13, 2025 · PreprintBioengineering