Improving diabetic retinopathy prognosis through HSV adjustments of OCTA images

• Main Author: Tan, James Hong Liang
• Other Authors: Liu Linbo
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/157750
• Institution: Nanyang Technological University

Diabetic Retinopathy (DR) is a common complication of diabetes that results from damage to the blood vessels at the rear of the eye. It is one of the leading causes of vision loss. Early detection and monitoring of DR using optical coherence tomography angiography (OCTA), through generation of three-dimensional images using back-reflected light to highlight moving blood cells, is important in the prognosis of patients with DR. OCTA is a preferred method over fluorescein angiography (FA) as it is non-invasive and safer, with further potential of enabling visualisation of finer capillaries due to its depth-resolved ability. One potential problem is that the probing beam may introduce projection artifacts in the deeper retinal layers when it passes through larger superficial vessels, affecting the quality of OCTA images produced and making it difficult to properly differentiate and identify blood vessels from surrounding tissues and structures in the eye. I will be exploring the idea of introducing colour models such as Red, Green and Blue (RBG), as well as Hue, Saturation and Value (HSV), into OCTA images to improve their visualisation. An experiment was carried out with a Graphical User Interface (GUI) that was designed and created using MATLAB to determine if there are specific HSV values that could be applied to OCTA images to not only suit the viewing preferences of clinicians of all ages and genders, but to also minimise errors in prognosis resulting from the presence of projection artifacts. Future studies incorporating multiple genetic and environmental variables, and merging more than one colour model may improve the results obtained in this paper to further improve the visualisation of OCTA images and accurate prognosis of DR.