Develop smartphone based portable polarisation sensitive eye imaging system in pre-clinical settings

Abstract

This Final Year Project (FYP) introduces a smartphone-based, portable polarisation-sensitive eye imaging system for pre-clinical applications. Integrating modern smartphone cameras with polarisation-sensitive optical coherence tomography (PS-OCT), the project aims to enhance ophthalmic diagnostics. A central feature is the 3D-printed platform, produced using the FlashForge Creator Pro. Despite its precision constraints, the platform effectively stabilises specimens for accurate PS-OCT imaging. Data collection yielded speckle patterns, A-Scans, and B-Scans, processed using MATLAB Python scripts for analysis. Min-max normalisation was applied across various imaging techniques to standardise data, but challenges arose in speckle data analysis due to inherent complexities. Collaboration with research associate involved using TensorFlow in a Neural Net Machine Learning Model to process B-Scans, with a 70/15/15 data split for training, validation, and testing. The SSIM analysis evaluated B-Scan quality before and after Machine Learning processing to assess accuracy. The findings lay groundwork for eye disease detection and management in animal models, and could one day potentially revolutionise early-stage eye disease detection and management for humans.