Related Experiment Video
Updated: Aug 1, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.8K
Deep Learning Approaches for Data Augmentation in Medical Imaging: A Review.
Aghiles Kebaili1, Jérôme Lapuyade-Lahorgue1, Su Ruan1
1Université Rouen Normandie, INSA Rouen Normandie, Université Le Havre Normandie, Normandie Univ, LITIS UR 4108, F-76000 Rouen, France.
Journal of Imaging
|April 27, 2023
Summary
Deep generative models like variational autoencoders, generative adversarial networks, and diffusion models can create realistic medical images for training. This approach addresses data scarcity and enhances deep learning performance in medical image analysis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Vision
Background:
- Limited medical training data hinders deep learning model development due to acquisition costs and privacy concerns.
- Traditional data augmentation methods often yield unconvincing and limited results for medical imaging tasks.
Purpose of the Study:
- To review deep generative models for medical image augmentation.
- To explore the potential of these models in enhancing deep learning for medical image analysis.
Main Methods:
- Focus on three key deep generative models: variational autoencoders (VAEs), generative adversarial networks (GANs), and diffusion models.
- Overview of the state-of-the-art for each model type in the context of medical image generation.
Main Results:
- These models generate diverse and realistic synthetic data, conforming to true data distributions.
- Potential applications include improving classification, segmentation, and cross-modal translation in medical imaging.
Conclusions:
- Deep generative models offer a promising solution to medical data scarcity.
- Further research is needed to fully leverage their capabilities for advancing medical image analysis.
