Synthetic RFMiD: Retinal Fundus Images Generated by Stable Diffusion

Synthetic RFMiD is a collection of synthetic retinal fundus images. The images were generated using a Stable Diffusion model. The dataset's size, authorship, and specific characteristics are not detailed in the provided metadata.

Use Cases

• Train or validate medical image classification models based on synthetic retinal fundus images.
• Augment existing real-world retinal image datasets with synthetic samples based on the described generation method.
• Benchmark the performance of generative models like Stable Diffusion in the medical imaging domain.
• Study the characteristics and potential biases of AI-generated medical imagery.

Strengths

• Images are generated by a Stable Diffusion model, a known generative AI technique.
• The dataset focuses on retinal fundus images, a specific and relevant medical imaging modality.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Collection Method

Images were generated using a Stable Diffusion model.