Fundus Photo Enhancement dataset
- Citation Author(s):
-
Hyungjin Chung (KAIST)Se Hie Park (Hallym University School of Medicine)Eui-Sang Chung (SNU Seoul Eye Clinic)Kayoung Yi (Hallym University School of Medicine)Jong Chul Ye (KAIST)
- Submitted by:
- Jong Chul Ye
- Last updated:
- DOI:
- 10.21227/hpd3-8d86
- Data Format:
139 views
- Categories:
- Keywords:
Abstract
We collected fundus photos from the Kangnam Sacred Heart Hospital, Hallym University School of Medicine, Seoul, South Korea (IRB approval number: 2022-10-026), that were obtained from 2000 to September 2022. The fundus photographs were taken by five skilled examiners using the KOWA Nonmyd 8S Fundus Camera (KOWA company, Japan). Among the 2,000 images, we chose 50 test images that were characterized as ``bad-quality'' due to one of the following reasons: media opacity, small pupil, or poor patient cooperation. For the rest of the 1,950 ``good-quality'' images, we removed the duplicates and those having sizes smaller than $256\times 256$. After the filtering, there are 1,152 images that are under the ``good-quality'' category.
Instructions:
We collected fundus photos from the Kangnam Sacred Heart Hospital, Hallym University School of Medicine, Seoul, South Korea (IRB approval number: 2022-10-026), that were obtained from 2000 to September 2022. The fundus photographs were taken by five skilled examiners using the KOWA Nonmyd 8S Fundus Camera (KOWA company, Japan). Among the 2,000 images, we chose 50 test images that were characterized as ``bad-quality'' due to one of the following reasons: media opacity, small pupil, or poor patient cooperation. For the rest of the 1,950 ``good-quality'' images, we removed the duplicates and those having sizes smaller than $256\times 256$. After the filtering, there are 1,152 images that are under the ``good-quality'' category.
Hello