HRF Image Database Viewer
Loading images...
Image Controls
Overview
This platform provides an interactive, high-performance viewer for the High-Resolution Fundus (HRF) Image Database. High-resolution retinal imaging is crucial for the early detection and management of ocular diseases. This tool is designed to assist researchers, clinicians, and students in visualizing and analyzing fundus images alongside their corresponding vessel segmentation maps and field of view (FOV) masks.
The HRF database is a vital resource for developing and validating algorithms for automated analysis of retinal images, particularly in the fields of diabetic retinopathy and glaucoma. By offering a seamless interface to explore these images, we aim to accelerate research, enhance medical training, and facilitate a deeper understanding of retinal pathology. The viewer is optimized for performance, utilizing lazy loading and modern web technologies to ensure a smooth experience even with high-resolution data.
How to Use
- Toggle Overlays: Use the 🎯 FOV Mask and 🩸 Vessel Overlay buttons to show or hide the corresponding layers on the images. The Field of View (FOV) mask highlights the usable area of the fundus image, while the vessel overlay displays the manually segmented retinal vasculature.
- Filter by Category: Use the dropdown menu to filter the gallery, showing only images from 'Healthy', 'Diabetic Retinopathy', or 'Glaucomatous' subjects. This is useful for comparative analysis.
- Sort Images: Organize the gallery by the original image 'Index' number or group them by 'Category' for a more structured view.
- View Full Size: Click on any thumbnail to open a high-resolution modal viewer. This allows for detailed inspection of the image and its active overlays.
- Keyboard Shortcuts: For efficiency, use keyboard shortcuts: 1 to toggle the FOV mask, 2 to toggle the vessel overlay, and Esc to close the full-size image viewer.
Performance Tips: The gallery uses a lazy-loading mechanism, meaning images are only loaded as they enter the viewport, which significantly speeds up initial page load. The progress bar at the top indicates the loading status of the visible images.
Future Directions
We are committed to the continuous improvement of this platform. Our roadmap includes several exciting enhancements to further empower the medical and research communities:
- Integration of More Datasets: We plan to incorporate other public fundus image datasets (e.g., DRIVE, STARE) to create a more comprehensive research hub.
- AI-Powered Analysis: Future versions will feature integrated AI models for real-time automated vessel segmentation, lesion detection (e.g., microaneurysms, hemorrhages), and diagnostic suggestions.
- Advanced Visualization: We are exploring techniques for 3D reconstruction of the retinal surface from stereo image pairs and creating dynamic visualizations to compare different segmentation algorithms.
- Collaborative Annotation Tools: A secure, web-based annotation tool will be developed to allow researchers to draw, label, and share their own segmentations or findings, fostering collaboration.
Your feedback and suggestions are invaluable in shaping the future of this resource. If you have ideas or would like to collaborate, please contact us.
About the HRF Image Database
The High-Resolution Fundus (HRF) Image Database comprises 45 fundus images: 15 images from healthy individuals, 15 from patients with diabetic retinopathy, and 15 from glaucomatous patients. Each image is accompanied by manually labeled vessel segmentation maps (gold standard) and field of view (FOV) masks.
For further details and to access the dataset, please visit the HRF Image Database.
Citation: If you use this resource for your research, please cite:
Budai, Attila; Bock, Rüdiger; Maier, Andreas; Hornegger, Joachim; Michelson, Georg. "Robust Vessel Segmentation in Fundus Images." International Journal of Biomedical Imaging, vol. 2013, 2013.