Back

Influence of hyperparameters on the performance of deep learning-based microrobotic localization under phantom tissue

Hoppe, J.; Nauber, R.; Castellanos Robles, D.; Czarske, J.; Medina Sanchez, M.

2025-08-16 bioengineering
10.1101/2025.08.13.670026 bioRxiv
Show abstract

For the effective operation of medical microrobots within living organisms and precise targeting, it is imperative to employ imaging techniques closely integrated with real-time deep-tissue tracking methods. However, due to a typically low Signal-to-Noise ratio images with strong background, it is hard for traditional tracking methods to achieve sufficient accuracy. This challenge can be addressed by deep learning-based tracking with a real-time detection model. However, a multitude of design choices and Hyperparameters influence the performance. In this study we compared the influence of the hyperparameters and model architecture versions of the "you only look once" (YOLO) network. We use experimental data from a magnetic microrobot imaged with Photoacoustics through 5 mm phantom tissue to evaluate the tracking in comparison with an optical reference. The deep-learning based methods consistently achieved lower missing-detection ratios. Regarding the Root Mean Square localization error, we observed that increasing the weight of the box loss function and utilizing the distribution focal loss can enhance the performance by 10%. Furthermore, it can be seen that YOLOv9 consistently outperformed its predecessor YOLOv8. This study quantifies the robustness of deep-learning based tracking of medical microrobots under tissues.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

1
Photoacoustics
11 papers in training set
Top 0.1%
12.2%
2
IEEE Transactions on Biomedical Engineering
38 papers in training set
Top 0.1%
8.3%
3
IEEE Transactions on Medical Imaging
18 papers in training set
Top 0.1%
8.1%
4
Biomedical Optics Express
84 papers in training set
Top 0.2%
7.1%
5
Scientific Reports
3102 papers in training set
Top 12%
7.1%
6
PLOS ONE
4510 papers in training set
Top 28%
6.3%
7
Optics Letters
13 papers in training set
Top 0.1%
6.3%
50% of probability mass above
8
Physics in Medicine & Biology
17 papers in training set
Top 0.1%
4.8%
9
Optics Express
23 papers in training set
Top 0.1%
3.8%
10
Journal of Neural Engineering
197 papers in training set
Top 1%
1.7%
11
Journal of Biomedical Optics
25 papers in training set
Top 0.4%
1.7%
12
Journal of Biophotonics
16 papers in training set
Top 0.3%
1.7%
13
Nature Communications
4913 papers in training set
Top 52%
1.6%
14
Frontiers in Bioengineering and Biotechnology
88 papers in training set
Top 2%
1.3%
15
IEEE Access
31 papers in training set
Top 0.6%
1.2%
16
Magnetic Resonance in Medicine
72 papers in training set
Top 0.5%
0.9%
17
NeuroImage
813 papers in training set
Top 5%
0.9%
18
PLOS Computational Biology
1633 papers in training set
Top 22%
0.9%
19
Computers in Biology and Medicine
120 papers in training set
Top 4%
0.9%
20
Sensors
39 papers in training set
Top 2%
0.9%
21
Frontiers in Neuroscience
223 papers in training set
Top 7%
0.8%
22
Medical Physics
14 papers in training set
Top 0.6%
0.8%
23
Journal of The Royal Society Interface
189 papers in training set
Top 4%
0.8%
24
Optica
25 papers in training set
Top 0.8%
0.7%
25
APL Bioengineering
18 papers in training set
Top 0.4%
0.7%
26
Communications Biology
886 papers in training set
Top 24%
0.7%
27
Annals of Biomedical Engineering
34 papers in training set
Top 1%
0.7%
28
Ultrasound in Medicine & Biology
10 papers in training set
Top 0.7%
0.6%
29
Advanced Science
249 papers in training set
Top 22%
0.6%
30
Human Brain Mapping
295 papers in training set
Top 5%
0.6%