DeepDRiD: Diabetic Retinopathy—Grading and Image Quality Estimation Challenge

Ruhan Liu; Xiangning Wang; Qiang Wu; Ling Dai; Xi Fang; Tao Yan; Jaemin Son; Shiqi Tang; Jiang Li; Zijian Gao; Adrian Galdran; J.M. Poorneshwaran; Hao Liu; Jie Wang; Yerui Chen; Prasanna Porwal; Gavin Siew Wei Tan; Xiaokang Yang; Chao Dai; Haitao Song; Mingang Chen; Huating Li; Weiping Jia; Dinggang Shen; Bin Sheng; Ping Zhang

doi:10.1016/j.patter.2022.100512

DeepDRiD: Diabetic Retinopathy—Grading and Image Quality Estimation Challenge

Ruhan Liu, Xiangning Wang, Qiang Wu, Ling Dai, Xi Fang, Tao Yan, Jaemin Son, Shiqi Tang, Jiang Li, Zijian Gao, Adrian Galdran, J.M. Poorneshwaran, Hao Liu, Jie Wang, Yerui Chen, Prasanna Porwal, Gavin Siew Wei Tan, Xiaokang Yang, Chao Dai, Haitao SongMingang Chen, Huating Li^*, Weiping Jia, Dinggang Shen^*, Bin Sheng^*, Ping Zhang

^*Corresponding author for this work

Department of Mathematics

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

167 Citations (Scopus)

143 Downloads (CityUHK Scholars)

Abstract

We described a challenge named “Diabetic Retinopathy (DR)—Grading and Image Quality Estimation Challenge” in conjunction with ISBI 2020 to hold three sub-challenges and develop deep learning models for DR image assessment and grading. The scientific community responded positively to the challenge, with 34 submissions from 574 registrations. In the challenge, we provided the DeepDRiD dataset containing 2,000 regular DR images (500 patients) and 256 ultra-widefield images (128 patients), both having DR quality and grading annotations. We discussed details of the top 3 algorithms in each sub-challenges. The weighted kappa for DR grading ranged from 0.93 to 0.82, and the accuracy for image quality evaluation ranged from 0.70 to 0.65. The results showed that image quality assessment can be used as a further target for exploration. We also have released the DeepDRiD dataset on GitHub to help develop automatic systems and improve human judgment in DR screening and diagnosis.

Original language	English
Article number	100512
Journal	Patterns
Volume	3
Issue number	6
Online published	20 May 2022
DOIs	https://doi.org/10.1016/j.patter.2022.100512
Publication status	Published - 10 Jun 2022

Research Keywords

artificial intelligence
challenge
deep learning
diabetic retinopathy
DSML2: Proof-of-concept Data science output has been formulated, implemented, and tested for one domain/problem
fundus image
image quality analysis
retinal image
screening
ultra-widefield

Publisher's Copyright Statement

This full text is made available under CC-BY 4.0. https://creativecommons.org/licenses/by/4.0/

Access Output

10.1016/j.patter.2022.100512

109600659.pdfFinal Published version, 1.84 MBLicence: CC BY-NC-ND

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

Liu, R., Wang, X., Wu, Q., Dai, L., Fang, X., Yan, T., Son, J., Tang, S., Li, J., Gao, Z., Galdran, A., Poorneshwaran, J. M., Liu, H., Wang, J., Chen, Y., Porwal, P., Wei Tan, G. S., Yang, X., Dai, C., ... Zhang, P. (2022). DeepDRiD: Diabetic Retinopathy—Grading and Image Quality Estimation Challenge. Patterns, 3(6), Article 100512. https://doi.org/10.1016/j.patter.2022.100512

@article{53a7212a874549378487b275552d787e,

title = "DeepDRiD: Diabetic Retinopathy—Grading and Image Quality Estimation Challenge",

abstract = "We described a challenge named “Diabetic Retinopathy (DR)—Grading and Image Quality Estimation Challenge” in conjunction with ISBI 2020 to hold three sub-challenges and develop deep learning models for DR image assessment and grading. The scientific community responded positively to the challenge, with 34 submissions from 574 registrations. In the challenge, we provided the DeepDRiD dataset containing 2,000 regular DR images (500 patients) and 256 ultra-widefield images (128 patients), both having DR quality and grading annotations. We discussed details of the top 3 algorithms in each sub-challenges. The weighted kappa for DR grading ranged from 0.93 to 0.82, and the accuracy for image quality evaluation ranged from 0.70 to 0.65. The results showed that image quality assessment can be used as a further target for exploration. We also have released the DeepDRiD dataset on GitHub to help develop automatic systems and improve human judgment in DR screening and diagnosis.",

keywords = "artificial intelligence, challenge, deep learning, diabetic retinopathy, DSML2: Proof-of-concept Data science output has been formulated, implemented, and tested for one domain/problem, fundus image, image quality analysis, retinal image, screening, ultra-widefield",

author = "Ruhan Liu and Xiangning Wang and Qiang Wu and Ling Dai and Xi Fang and Tao Yan and Jaemin Son and Shiqi Tang and Jiang Li and Zijian Gao and Adrian Galdran and J.M. Poorneshwaran and Hao Liu and Jie Wang and Yerui Chen and Prasanna Porwal and {Wei Tan}, {Gavin Siew} and Xiaokang Yang and Chao Dai and Haitao Song and Mingang Chen and Huating Li and Weiping Jia and Dinggang Shen and Bin Sheng and Ping Zhang",

year = "2022",

month = jun,

day = "10",

doi = "10.1016/j.patter.2022.100512",

language = "English",

volume = "3",

journal = "Patterns",

issn = "2666-3899",

publisher = "Cell Press",

number = "6",

}

Liu, R, Wang, X, Wu, Q, Dai, L, Fang, X, Yan, T, Son, J, Tang, S, Li, J, Gao, Z, Galdran, A, Poorneshwaran, JM, Liu, H, Wang, J, Chen, Y, Porwal, P, Wei Tan, GS, Yang, X, Dai, C, Song, H, Chen, M, Li, H, Jia, W, Shen, D, Sheng, B & Zhang, P 2022, 'DeepDRiD: Diabetic Retinopathy—Grading and Image Quality Estimation Challenge', Patterns, vol. 3, no. 6, 100512. https://doi.org/10.1016/j.patter.2022.100512

TY - JOUR

T1 - DeepDRiD

T2 - Diabetic Retinopathy—Grading and Image Quality Estimation Challenge

AU - Liu, Ruhan

AU - Wang, Xiangning

AU - Wu, Qiang

AU - Dai, Ling

AU - Fang, Xi

AU - Yan, Tao

AU - Son, Jaemin

AU - Tang, Shiqi

AU - Li, Jiang

AU - Gao, Zijian

AU - Galdran, Adrian

AU - Poorneshwaran, J.M.

AU - Liu, Hao

AU - Wang, Jie

AU - Chen, Yerui

AU - Porwal, Prasanna

AU - Wei Tan, Gavin Siew

AU - Yang, Xiaokang

AU - Dai, Chao

AU - Song, Haitao

AU - Chen, Mingang

AU - Li, Huating

AU - Jia, Weiping

AU - Shen, Dinggang

AU - Sheng, Bin

AU - Zhang, Ping

PY - 2022/6/10

Y1 - 2022/6/10

N2 - We described a challenge named “Diabetic Retinopathy (DR)—Grading and Image Quality Estimation Challenge” in conjunction with ISBI 2020 to hold three sub-challenges and develop deep learning models for DR image assessment and grading. The scientific community responded positively to the challenge, with 34 submissions from 574 registrations. In the challenge, we provided the DeepDRiD dataset containing 2,000 regular DR images (500 patients) and 256 ultra-widefield images (128 patients), both having DR quality and grading annotations. We discussed details of the top 3 algorithms in each sub-challenges. The weighted kappa for DR grading ranged from 0.93 to 0.82, and the accuracy for image quality evaluation ranged from 0.70 to 0.65. The results showed that image quality assessment can be used as a further target for exploration. We also have released the DeepDRiD dataset on GitHub to help develop automatic systems and improve human judgment in DR screening and diagnosis.

AB - We described a challenge named “Diabetic Retinopathy (DR)—Grading and Image Quality Estimation Challenge” in conjunction with ISBI 2020 to hold three sub-challenges and develop deep learning models for DR image assessment and grading. The scientific community responded positively to the challenge, with 34 submissions from 574 registrations. In the challenge, we provided the DeepDRiD dataset containing 2,000 regular DR images (500 patients) and 256 ultra-widefield images (128 patients), both having DR quality and grading annotations. We discussed details of the top 3 algorithms in each sub-challenges. The weighted kappa for DR grading ranged from 0.93 to 0.82, and the accuracy for image quality evaluation ranged from 0.70 to 0.65. The results showed that image quality assessment can be used as a further target for exploration. We also have released the DeepDRiD dataset on GitHub to help develop automatic systems and improve human judgment in DR screening and diagnosis.

KW - artificial intelligence

KW - challenge

KW - deep learning

KW - diabetic retinopathy

KW - DSML2: Proof-of-concept Data science output has been formulated, implemented, and tested for one domain/problem

KW - fundus image

KW - image quality analysis

KW - retinal image

KW - screening

KW - ultra-widefield

UR - http://www.scopus.com/inward/record.url?scp=85131696394&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85131696394&origin=recordpage

U2 - 10.1016/j.patter.2022.100512

DO - 10.1016/j.patter.2022.100512

M3 - RGC 21 - Publication in refereed journal

C2 - 35755875

SN - 2666-3899

VL - 3

JO - Patterns

JF - Patterns

IS - 6

M1 - 100512

ER -

DeepDRiD: Diabetic Retinopathy—Grading and Image Quality Estimation Challenge

Abstract

Research Keywords

Publisher's Copyright Statement

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this