Annotated Infant Fundus Images for Object Detection of Key Structures (Optic Disk and Macula)

Published: 16 November 2021| Version 3 | DOI: 10.17632/9kvk7nkhx6.3
Yinsheng Zhang


The data set contains fundus images for object detection research. Some images have clipped corners (patient info watermark is printed in this area) to ensure they are de-identified. All images are annotated by one senior pediatric ophthalmologist in SEH (Shenzhen Eye Hospital). An image annotation tool is developed based on the "VGG Image Annotator (VIA)" tool from the University of Oxford.The tool is included in this repository. The optic disk and macula in each fundus image are annotated by rectangular bounding boxes. This repository also contains related materials for our case study on this data set.



Computer Vision, Fundus Imaging