Related Experiment Video
Updated: Oct 22, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.0K
Dataset Growth in Medical Image Analysis Research
1School of Electrical Engineering, Tel Aviv University, Tel Aviv 69978, Israel.
Journal of Imaging
|August 30, 2021
Summary
Medical image analysis researchers face data scarcity. Datasets are growing exponentially, with MRI, CT, and fMRI data sizes increasing annually, impacting research standards.
Area of Science:
- Medical image analysis
- Computational imaging
- Data science in healthcare
Background:
- Medical image analysis research relies on large datasets.
- Researchers often face challenges in accessing sufficient data, termed 'data starved'.
- Evolving community standards may necessitate the use of increasingly larger datasets.
Purpose of the Study:
- To analyze the growth trends of medical image dataset sizes used in research.
- To forecast future dataset size requirements in medical image analysis.
- To validate growth predictions with recent conference data.
Main Methods:
- Scanned MICCAI conference proceedings (2011-2018) to identify papers using human MRI, CT, or fMRI datasets.
- Extracted and analyzed dataset sizes, calculating median and geometric mean growth rates.
- Forecasted dataset sizes for MICCAI 2019 and validated with actual data from the conference.
Main Results:
- Median dataset sizes increased 3-10 fold between 2011 and 2018.
- Exponential growth observed in geometric mean dataset size: 21% (MRI), 24% (CT), 31% (fMRI) annually (Phase I).
- MICCAI 2019 data confirmed forecasts, with revised annual growth rates of 27% (MRI), 30% (CT), 32% (fMRI).
Conclusions:
- Medical image dataset sizes are growing exponentially, driven by implicit community standards.
- Accurate predictions of dataset size growth can be made for future conferences.
- These findings have implications for data management and resource allocation in medical imaging research.
