Towards multi-center glaucoma OCT image screening with semi-supervised joint structure and function multi-task learning.

•It is the first study on unifying structure analysis and function regression for glaucoma screening from OCT images.•The semi-supervised smoothness assumption is made to solve the missing regression label problem.•A multi-task learning network is proposed to explore the structure-function relationship for glaucoma screening.•Extensive experiments on large-scale multi-center datasets demonstrate the effectiveness of the multi-task learning model.