Local Estimation of the Degree of Optic Disc Swelling from Color Fundus Photography

An Integrated Fundus Image Segmentation Algorithm for Multiple Eye Ailments

“…Typically, when there is some ailment in an eye: either there are red spots [12], yellow circles [13], white spots, or cotton type blurry objects in the eye [14]. There might be a change in the size of the optical disc [15], a change in the number and dimensions of veins/arteries in the eyes, or simply an abnormal variation of texture in the retina area of the cornea area [15]. Hence, the objective of segmentation algorithms is to extract and segment those areas that give a clear-cut indication of a problem in the eye and clinical evidence that there is a specific type of eye disease.…”

“…By segmenting different parts of the image, more precise pathology can be done and it will assure image based evidence care for ocular and medical disorders. The form, thickness, and texture of the eye regions aid in the analysis of the overall eye health [15][16][17][18]. If the shape, colour, and texture of the various parts of the eye (arteries, optical disc, eyeball size, and veins) deviate from the normal pattern of growth, the medical professional will be able to diagnose an eye condition [19].…”

Estimation of volumetric optical coherence tomography measurements from 2D color fundus photographs using machine learning

“…The main contribution of this thesis is the innovative application of machine learning methods for assessing the severity, in both a qualitative and quantitative manner, of optic disc edema. In Chapter 3 we presented a traditional machine learning approach in the form of a random forest ensemble regression model to estimate thickness between retinal layers which was published as a conference paper in the MICCAI conference's Ophthalmic Medical Image Analysis workshop [41] in 2018 and was also given as an oral presentation. For comparison, and due to its recent success in both image classification and medical image analysis, a deep learning method was implemented on the same data used for the random forest, which is described in Chapter 4 and was accepted as a conference abstract at the ARVO conference in 2019 [45].…”

“…Note: A version of this work has previously been published and presented at the MICCAI conference's Ophthalmic Medical Image Analysis workshop in 2018 [41].…”