Physics-guided image segmentation and classification for tear film break-ups (Hangjie Ji)

Prerequisites: Multivariable calculus, differential equations, and basic programming skills.

Outline: Human eyes are coated with a multi-layer liquid film (a tear film) that establishes rapidly after a blink. Tear film thinning and break-up during interblink periods play a crucial role in Dry Eye Syndrome, which causes blurred vision, tearing, and inflammation of the ocular surface (1). Tear film dynamics can be modeled by nonlinear PDEs (2). To facilitate personalized diagnosis, biomedical image segmentation and classification are needed to predict the dynamics of tear film break-ups. In this project, we will develop a physics-guided image segmentation framework to identify and classify break-up spots in tear film images.

Research objectives: Exploration of human tear film image segmentation and classification guided by PDE models to better understand the spatial instability of tear film dynamics. The segmentation results will be further analyzed to estimate the influence of key factors that lead to tear film break-ups, such as evaporation, osmolarity dynamics, and permeability conditions.

Outcomes: A biomedical image segmentation and classification framework guided by PDE models for human tear film dynamics. Results will provide tools to quickly predict the development of tear film break-ups and estimate key parameters for the PDE models.

References:

  • Johnson M, Murphy P. Changes in the tear film and ocular surface from dry eye syndrome. Prog Ret Eye Res. 2004;23:449-74.
  • Braun R. Dynamics of the tear film. Ann Rev Fluid Mech. 2012;44:267–297.