Artificial Intelligence: Applications for Pathologies of the Anterior Segment of the Eye

Dr Damien Gatinel

Meeting:  2022 RANZCO


Date:      -


Session Time:      -

Synopsis: The use of artificial intelligence (AI) pro- vides a better understanding of data collected through biometric and imaging techniques. It can increase the accuracy or the effectiveness of solutions intended to correct certain optical defects of the eye caused by struc- tural alterations responsible for a reduction in the trans- parency of the ocular media and/or refractive errors.
This work has been mainly applied to the study of cor- neal topography and imaging, description of the ocular wavefront, ocular biometry and improving the accuracy of the power calculation of intraocular lenses.
We used AI techniques to carry out studies to character- ise the alterations observed during the evolution of keratoconus, as well as the spatial distribution of corneal oedema through the use of neural networks to tomo- graphic image analysis.
Using artificial intelligence and the results from the description of the ocular wavefront on a new basis, we were able to estimate the real impact of high degree aber- rations on subjective refraction.
From a thick lens paraxial eye model, we established an analytical formula to calculate the position of the princi- pal planes of an implant based on its geometry and the main biometric constants. This work made it possible to train an algorithm to predict the effective position of an implant, taking into account its geometry and esta- blishing an original biometric calculation formula com- bining principles of paraxial optics and artificial intelligence algorithms applied to an eye model made up of thick lenses.