Evaluating Utility of
Memory Efficient
Medical Image Generation
A Study on Lung Nodule Segmentation
This paper evaluates quality and effectiveness of synthetic data by testing its impact on downstream segmentation tasks. The results show:
- Segmentation models trained solely on synthetic data perform on par compared to training on RWD
- Augmenting RWD with synthetic images improves model performance significantly
Ask us anything about Synthetic Data
EXTRAs
CASE STUDY
Synthetic Training Data Improves Lung Nodule Classification
This case study shows how synthetic 3D Lung CTs including nodules of different size and texture can be used to enhance a best-in-class classification model.