Hyperparameter selection for dataset‐constrained semantic segmentation: Practical machine learning optimization
Journal of Applied Clinical Medical Physics · 2024
DOI: 10.1002/acm2.14542 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Applied Clinical Medical Physics, 2024. Verfasst von Chris Boyd und 6 weiteren Autor:innen. Beteiligte Einrichtung: Allied Health and Human Performance, University of South Australia, Adelaide, Australia. Hauptthemen laut MEDLINE: Machine Learning, Tomography, X-Ray Computed, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Purpose/aim This paper provides a pedagogical example for systematic machine learning optimization in small dataset image segmentation, emphasizing hyperparameter selections. A simple process is presented for medical physicists to examine hyperparameter optimization. This is also applied to a case‐study, demonstrating the benefit of the method. Materials and methods An unrestricted public Computed Tomography (CT) dataset, with binary organ segmentation, was used to develop a multiclass segmentation model. To start the optimization process, a preliminary manual search of hyperparameters was conducted and from there a grid search identified the most influential result metrics. A total of 658 different models were trained in 2100 h, using 13 160 effective patients. The quantity of results was analyzed using random forest regression, identifying relative hyperparameter impact.
Results
Metric implied segmentation quality (accuracy 96.8%, precision 95.1%) and visual inspection were found to be mismatched. In this work batch normalization was most important, but performance varied with hyperparameters and metrics selected. Targeted grid‐search optimization and random forest analysis of relative hyperparameter importance, was an easily implementable sensitivity analysis approach.
Conclusion
The proposed optimization method gives a systematic and quantitative approach to something intuitively understood, that hyperparameters change model performance. Even just grid search optimization with random forest analysis presented here can be informative within hardware and data quality/availability limitations, adding confidence to model validity and minimize decision‐making risks. By providing a guided methodology, this work helps medical physicists to improve their model optimization, irrespective of specific challenges posed by datasets and model design.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Allied Health and Human Performance, University of South Australia, Adelaide, Australia.
- Medical Physics and Radiation Safety, South Australia Medical Imaging, Adelaide, Australia.
- Department of Neurology, Royal Adelaide Hospital, Adelaide, Australia.
- Adelaide Medical School, The University of Adelaide, Adelaide, Australia.
- Discipline of Surgery, University of Adelaide, Adelaide, Australia.
- Department of Vascular and Endovascular Surgery, Royal Adelaide Hospital, Adelaide, Australia.
ORCID-Profile: Chris Boyd
Zitieren
APA
Boyd, C., Brown, G. C., Kleinig, T. J., Mayer, W., Dawson, J., Jenkinson, M., & Bezak, E. (2024). Hyperparameter selection for dataset‐constrained semantic segmentation: Practical machine learning optimization. Journal of Applied Clinical Medical Physics. https://doi.org/10.1002/acm2.14542
Vancouver
Boyd C, Brown GC, Kleinig TJ, Mayer W, Dawson J, Jenkinson M, et al. Hyperparameter selection for dataset‐constrained semantic segmentation: Practical machine learning optimization. Journal of Applied Clinical Medical Physics. 2024. doi:10.1002/acm2.14542
BibTeX
@article{Boyd2024,
author = {Boyd, Chris and Brown, Gregory C and Kleinig, Timothy J and Mayer, Wolfgang and Dawson, Joseph and Jenkinson, Mark and Bezak, Eva},
title = {Hyperparameter selection for dataset‐constrained semantic segmentation: Practical machine learning optimization},
journal = {Journal of Applied Clinical Medical Physics},
year = {2024},
doi = {10.1002/acm2.14542}
}
EUVIMED