EUVIMEDEUVIMED entdecken
Crossref

Using diffusion models to generate synthetic labeled data for medical image segmentation

Daniel G Saragih, Atsuhiro Hibi, Pascal N Tyrrell

International Journal of Computer Assisted Radiology and Surgery · 20.06.2024 · 19 (8) · 1615-1625

DOI: 10.1007/s11548-024-03213-z ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Computer Assisted Radiology and Surgery, online erschienen am 20.06.2024. Verfasst von Daniel G Saragih und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Medical Imaging, University of Toronto, 263 McCaul Street, Toronto, M5T 1W7, ON, Canada. Hauptthemen laut MEDLINE: Machine Learning. Die Arbeit wurde bisher 35-mal zitiert (Crossref) und nennt 29 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

PURPOSE: Medical image analysis has become a prominent area where machine learning has been applied. However, high-quality, publicly available data are limited either due to patient privacy laws or the time and cost required for experts to annotate images. In this retrospective study, we designed and evaluated a pipeline to generate synthetic labeled polyp images for augmenting medical image segmentation models with the aim of reducing this data scarcity. METHODS: We trained diffusion models on the HyperKvasir dataset, comprising 1000 images of polyps in the human GI tract from 2008 to 2016. Qualitative expert review, Fréchet Inception Distance (FID), and Multi-Scale Structural Similarity (MS-SSIM) were tested for evaluation. Additionally, various segmentation models were trained with the generated data and evaluated using Dice score (DS) and Intersection over Union (IoU). RESULTS: Our pipeline produced images more akin to real polyp images based on FID scores. Segmentation model performance also showed improvements over GAN methods when trained entirely, or partially, with synthetic data, despite requiring less compute for training. Moreover, the improvement persists when tested on different datasets, showcasing the transferability of the generated images. CONCLUSIONS: The proposed pipeline produced realistic image and mask pairs which could reduce the need for manual data annotation when performing a machine learning task. We support this use case by showing that the methods proposed in this study enhanced segmentation model performance, as measured by Dice and IoU scores, when trained fully or partially on synthetic data.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningHumansRetrospective StudiesImage Processing, Computer-Assisted

Stichwörter der Autor:innen

Data augmentationDiffusion modelsMachine learningPolyp image generation

Autor:innen und Institutionen

  • Department of Medical Imaging, University of Toronto, 263 McCaul Street, Toronto, M5T 1W7, ON, Canada.
  • Institute of Medical Science, University of Toronto, Toronto, ON, Canada.
  • Department of Medical Imaging, University of Toronto, 263 McCaul Street, Toronto, M5T 1W7, ON, Canada. pascal.tyrrell@utoronto.ca.
  • Institute of Medical Science, University of Toronto, Toronto, ON, Canada. pascal.tyrrell@utoronto.ca.
  • Department of Statistical Sciences, University of Toronto, Toronto, ON, Canada. pascal.tyrrell@utoronto.ca.

Zitieren

APA

Saragih, D. G., Hibi, A., & Tyrrell, P. N. (2024). Using diffusion models to generate synthetic labeled data for medical image segmentation. International Journal of Computer Assisted Radiology and Surgery, 19(8), 1615-1625. https://doi.org/10.1007/s11548-024-03213-z

Vancouver

Saragih DG, Hibi A, Tyrrell PN. Using diffusion models to generate synthetic labeled data for medical image segmentation. International Journal of Computer Assisted Radiology and Surgery. 2024;19(8):1615-1625. doi:10.1007/s11548-024-03213-z

BibTeX
@article{Saragih2024,
  author = {Saragih, Daniel G and Hibi, Atsuhiro and Tyrrell, Pascal N},
  title = {Using diffusion models to generate synthetic labeled data for medical image segmentation},
  journal = {International Journal of Computer Assisted Radiology and Surgery},
  year = {2024},
  volume = {19},
  number = {8},
  pages = {1615--1625},
  doi = {10.1007/s11548-024-03213-z}
}
Kostenlos starten →