Release of the world's largest annotated database of glaucoma fundus images by Prof Xiulan Zhang at MICCAI 2018
Resource: Clinical Research Center
Edited by Xianjing Wei
The 21st International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) was held in Granada, Spain, from September 16th to 20th, 2018. More than 1500 researchers major in medical imaging from all over the world gathered at this event. MICCAI covers various fields in medical imaging, including the development of post-processing algorithms for traditional medical images such as CT and MRI, as well as new directions such as optical imaging technology and artificial intelligence assisted image processing. Ophthalmology, as a professional field with many imaging methods, occupies a large space. Professor Xiulan Zhang and Dr Fei Li from our center were invited to participate the MICCAI conference and share their latest researches in glaucoma imaging.
At the Ophthalmic Medical Image Analysis (OMIA) event held on the 20th, Professor Xiulan Zhang and Yanwu Xu, Chief Architect of Medical AI, together released the major attraction of the OMIA, the REFUGE (REtinal FUndus Glaucoma ChallengE) database (https://refuge.grand-challenge.or. G/). The public database REFUGE is the most comprehensive database of glaucoma fundus images based on annotations from seven ophthalmologists. Before the conference, it has attracted much attention. 230 teams from more than 20 countries and regions all over the world signed up for the REFUGE Challenge to be held at the conference.
Xiulan Zhang and Yanwu Xu together released the REFUGE database
Dr. Fei Li was invited to give a speech at the conference to share his latest progress on visual field based automatic differentiation of visual field using deep neural network, which received the Best Paper Award at the OMIA conference.
Dr. Fei Li at the conference
As the Board Member of APGS and AACGC, Professor Xiulan Zhang is devoted to the most cutting edge topics of glaucoma researches. Currently, application of artificial intelligence technology in clinical diagnosis of glaucoma has become very popular. REFUGE is a database based on fundus photography, which can be used for training and comparison of various neural network algorithms.
Besides fundus photography, Professor Zhang's team in cooperation with Professor Yu Qiao’s team from Shenzhen Institute of Advanced Technologies, CAS have done a lot of research on visual field and OCT interpretation using neural network.
Professor Zhang believes that the future belongs to big data and artificial intelligence, so sharing existing data resources is very important. The More researchers having access to high-quality data, the better algorithms will be developed. Although many products and articles claim to have high diagnostic accuracy for glaucoma, there is no database for fine annotations of key structures in fundus photography. For this reason, Professor Zhang and Xu launched REFUGE database, which filled this gap. Besides diagnosis, it can be used to train the neural network to process complex tasks and make the most accurate analysis based on imaging details. It will be the powerful assistant of glaucoma clinical diagnosis and contribute to the development of medical artificial intelligence in China.