You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

Glaucoma Detection Dataset

A dataset of annotated fundus images for developing glaucoma detection models. This dataset is a private data that not available for public.



Dataset Description

This dataset consists of labeled fundus images classified into two categories: glaucoma and non-glaucoma. Each image includes annotations with polygonal shapes marking the disc and cup regions. But, not all of the fundus image is provided with cup annotation.

Data Structure

  • Folder Structure:
    raw_image
    β”œβ”€β”€ raw/
    β”‚   β”œβ”€β”€ fundus/
    β”‚   β”œβ”€β”€ mask/
    β”œβ”€β”€ filtered/
    β”‚   β”œβ”€β”€ fundus/
    β”‚   β”œβ”€β”€ mask/
    β”œβ”€β”€ metadata/
    β”‚   β”œβ”€β”€ glaucoma_annotations.json/
    β”‚   β”œβ”€β”€ imgs_filtered_info.csv/
    β”‚   β”œβ”€β”€ naming_convention.md/
    β”‚   └── non_glaucoma_annotations.json/
    └── README.md
    

Dataset Summary

The OCT image is not provided to maintain patient confidentiality

Label OCT Fundus Total Image Patient
Non-glaucoma 60 197 257 41
Glaucoma 79 222 301 50
Total 139 419 558 91

The annotation information:

  • Disc label dot β†’ 30
  • Cup label dot β†’ 20

Author & Contributor

  • Anindita Septiarini β†’ Project Lead
  • Nur Khomah Fatmawati β†’ Data Collection
  • Bugi Sulistiyo β†’ Annotator & Data Preprocessing
  • Eko Rahmat Darmawan β†’ Annotator
Downloads last month
31