Diabetic retinopathy is one of the complications of diabetes leading to blindness. Early detection through periodic screening of retinal images can help reduce the long-term effects and blindness by as much as 90%. Its resource demands, however are high requiring eye doctors and equipment. This is particularly problematic in remote and resource-poor areas where the detection is most needed. With half billion people already living with diabetes worldwide, and their number growing at a rapid pace due to rapid industrialization, unhealthy diets and sedentary lifestyles, innovative technology-enabled solutions are essential.
Imagine if artificial intelligence (AI) technologies, in particular deep learning, could offer a way to implement real-time screening at relatively low cost and take retinal screening to the diabetic patient in primary care. In conjunction with the Sydney Informatics Hub, you are invited to assist Assoc. Prof. Ravi Seethamraju from the University of Sydney Business School and Prof. Krishna Sundar of the Digital Innovation Lab at the Indian Institute of Management Bangalore, in designing a low-cost solution.
The challenge is to develop a model that could classify diabetic retinopathy images into five categories [Proliferative DR (PDR), Normal Non-proliferative DR (NPDR), mild NPDR, moderate NPDR and severe NPDR] for a given dataset of images. You will be free to use any publicly available data and deep learning/machine learning algorithm (or any automated method) of your choice for building, training and validating your model.
1st place (DR Prize) – $7500: Simon Cai
2nd place (Kaggle Prize) – $2000: Mike Li
3rd place (Artemis Prize) – $500: Patrick Hao and Andrew Lee