CornealConfocalDataset: Multi-Disease In Vivo Microscopy with Pixel Annotations

A 178 MB dataset of in vivo confocal microscopy images for corneal disease analysis. It contains pixel-level expert annotations for multiple diseases, supporting machine learning model development. The dataset is the official repository for a corresponding research paper.

Use Cases

• Training segmentation models based on pixel-level expert annotations.
• Developing classifiers for multi-disease corneal diagnosis from confocal images.
• Benchmarking computer vision algorithms for medical image analysis.
• Studying cellular and structural features of diseased corneal tissue.

Strengths

• Contains pixel-level expert annotations, a high-quality labeling standard.
• Specifically designed for multi-disease analysis, increasing utility.
• Released under a permissive CC-BY-4.0 license for reuse.
• Associated with a formal research paper, indicating academic rigor.

Limitations

• The exact number of images and rows is not specified in the metadata.
• Column names and a detailed data dictionary are not provided.
• The dataset is packaged as a single ZIP file without internal structure details.

Provenance

Source

harvey mu, associated with a research paper titled 'A multi-disease in vivo confocal microscopy dataset with pixel-level expert annotations for corneal disease analysis'.

Collection Method

In vivo confocal microscopy imaging with subsequent expert annotation.

Time Range

null

Freshness

Last updated on 2026-03-24.

Geography

null