IDRiD: Diabetic Retinopathy – Segmentation and Grading Challenge

TLDR: The set-up and results of this challenge that is primarily based on Indian Diabetic Retinopathy Image Dataset (IDRiD), which received a positive response from the scientific community, have the potential to enable new developments in retinal image analysis and image-based DR screening in particular.

About: This article is published in Medical Image Analysis.The article was published on 2020-01-01 and is currently open access. It has received 169 citations till now. The article focuses on the topics: Population.

Figures

Table 6. Summary of data augmentation, normalization and pre-processing in the competing solutions. Where, RF, RR, RS, RT, RC represent random flip, rotation, scaling, translation and crop respectively.

Fig. 10. Boxplots (a,b) showing dispersion of Euclidean distance for individual methods for OD and fovea and (c) showing the dispersion of Jaccard index for OD segmentation task. Boxplots show quartile ranges of the scores on the test dataset; plus sign indicate outliers (full range of data is not shown).

Fig. 2. Retinal photograph and different annotations: (a) sample fundus image from the IDRiD dataset; sample ground truths of (b-f) MAs, HEs, SEs, EXs and OD respectively.

Table 8. Sub-challenge – 1 “Off-site” leaderboard highlighting top 4 teams from each lesion (MAs, HEs, SEs and EXs) segmentation task on the testing dataset. It details the approach followed by respective team and external dataset used for training their model (if any).

Fig. 5. Architecture for dilation block.

Table A.2. Comparison of different databases with the IDRiD database

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Idrid: diabetic retinopathy -" ?

The recent scientific advances in computing capacity and machine learning approaches provide an avenue for biomedical scientists to reach this goal. In this paper, the authors report the set-up and results of this challenge that is primarily based on Indian Diabetic Retinopathy Image Dataset ( IDRiD ). This paper outlines the challenge, its organization, the dataset used, evaluation methods and results of top performing participating solutions. The authors observe that the top performing approaches utilize a blend of clinical information, data augmentation, and the ensemble of models. Org/open-access/indian-diabetic-retinopathy-imagedataset-idrid IDRiD: Diabetic Retinopathy Segmentation and Grading Challenge Prasanna Porwala, b,1, ∗, Samiksha Pachadea, b,1, Manesh Kokarea,1, Girish Deshmukhc,1, Jaemin Sond, Woong Baed, Lihong Liue, Jianzong Wange, Xinhui Liue, Liangxin Gaoe, TianBo Wue, Jing Xiaoe, Fengyan Wangf, Baocai Yinf, Yunzhi Wangg, Gopichandh Danalag, Linsheng Heg, Yoon Ho Choih, Yeong Chan Leeh, Sang Hyuk Jungh, Zhongyu Lii, Xiaodan Suij, Junyan Wul, Xiaolong Lim, Ting Zhoun, János Tótho, Agnes Barano, Avinash Korip, Varghese Alexp, Sai Saketh Chennamsettyp, Mohammed Safwanp, Xingzheng Lyuq, r, Li Chengr, Qinhao Chus, Pengcheng Lis, Xin Jit, Sanyuan Zhangq, Yaxin Shenu, v, Ling Daiu, v, Oindrila Sahax, Rachana Sathishx, Tânia Meloy, Teresa Araújoy, z, Balázs Harangio, Bin Shengu, v, Ruogu Fangw, Debdoot Sheetx, Andras Hajduo, Yuanjie Zhengj, Ana Maria Mendonçay, z, Shaoting Zhangi, Aurélio Campilhoy, z, Bin Zhengg, Dinggang Shenk, Luca Giancardob,1, Gwenolé Quellecaa,1, Fabrice Mériaudeauab, ac,1 aShri Guru Gobind Singhji Institute of Engineering and Technology, Nanded, India bSchool of Biomedical Informatics, The University of Texas Health Science Center at Houston, USA cEye Clinic, Sushrusha Hospital, Nanded, Maharashtra, India dVUNO Inc., Seoul, Republic of Korea ePing An Technology ( Shenzhen ) Co., Ltd, China fiFLYTEK Research, Hefei, China gSchool of Electrical and Computer Engineering, University of Oklahoma, USA hSamsung Advanced Institute for Health Sciences & Technology ( SAIHST ), Sungkyunkwan University, Seoul, Republic of Korea iDepartment of Computer Science, University of North Carolina at Charlotte, USA jSchool of Information Science and Engineering, Shandong Normal University, China kDepartment of Radiology and BRIC, the University of North Carolina at Chapel Hill, USA lCleerly Inc., New York, United States mVirginia Tech, Virginia, United States nUniversity at Buffalo, New York, United States oUniversity of Debrecen, Faculty of Informatics 4002 Debrecen, POB 400, Hungary pIndividual Researcher, India qCollege of Computer Science and Technology, Zhejiang University, Hangzhou, China rMachine Learning For Bioimage Analysis Group, Bioinformatics Institute, A * STAR, Singapore sSchool of Computing, National University of Singapore, Singapore tBeijing Shanggong Medical Technology Co., Ltd., China uDepartment of Computer Science and Engineering, Shanghai Jiao Tong University, China vMoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University, China wJ. Crayton Pruitt Family Department of Biomedical Engineering, University of Florida, USA xIndian Institute of Technology Kharagpur, India yINESC TEC Institute for Systems and Computer Engineering, Technology and Science, Porto, Portugal zFEUP Faculty of Engineering of the University of Porto, Porto, Portugal aaINSERM, UMR 1101, Brest, France abDepartment of Electrical and Electronic Engineering, Universiti Teknologi PETRONAS, Malaysia acImViA/IFTIM, Université de Bourgogne, Dijon, France ∗Corresponding author Email address: porwal. All others contributed results of their algorithm ( s ) presented in the paper Preprint submitted to Medical Image Analysis June 10, 2019 Manuscript Click here to download Manuscript: elsarticle-template-V6. These findings have the potential to enable new developments in retinal image analysis and image-based DR screening in particular.