Semi-automated identification of cones in the human retina using circle Hough transform

Abstract

A large number of human retinal diseases are characterized by a progressive loss of cones, the photoreceptors critical for visual acuity and color perception. Adaptive Optics (AO) imaging presents a potential method to study these cells in vivo. However, AO imaging in ophthalmology is a relatively new phenomenon and quantitative analysis of these images remains difficult and tedious using manual methods. This paper illustrates a novel semi-automated quantitative technique enabling registration of AO images to macular landmarks, cone counting and its radius quantification at specified distances from the foveal center. The new cone counting approach employs the circle Hough transform (cHT) and is compared to automated counting methods, as well as arbitrated manual cone identification. We explore the impact of varying the circle detection parameter on the validity of cHT cone counting and discuss the potential role of using this algorithm in detecting both cones and rods separately.

Cone Photoreceptor Cell Segmentation and Diameter Measurement on Adaptive Optics Images Using Circularly Constrained Active Contour Model

“… 1 – 3 Quantitative assessment of the mosaic through metrics on AO retinal images, such as cone density and spacing, has shown potential for clinical application 2 , 4 with substantial efforts already realized toward assembling normative databases. 5 – 12 To overcome the tedious task of manually identifying cones and to remove the variability of human graders, various automated algorithms have been developed for two types of AO modalities: confocal 13 – 16 and nonconfocal 17 – 19 AO scanning light ophthalmoscopy (AOSLO). However, most quantitative metrics have been based on representing each cone as a point.…”

Cellular imaging of inherited retinal diseases using adaptive optics

“…Aside from difficulties in accurate cone identification, the impracticality of manual image grading by observers and its poor repeatability can also render metric applications as unreliable [95]. Recent efforts to produce automated analytic tools for AO images have shown promise in both confocal and non-confocal settings [119,120], with the latter being used in achromatopsia and Stargardt disease [121,122]. Although further work is needed to characterise the performance of these algorithms in relation to different metrics [123], this AO-FIO AOSLO Canal-like foveal schisis cavities seen, with a spoke wheel pattern of inner retinal folds.…”

Photoreceptor-Based Biomarkers in AOSLO Retinal Imaging

“… 114 For cone identification, many groups have developed methods to detect cones automatically in confocal AOSLO images. 65 , 69 – 71 , 159 , 162 , 163 Garrioch et al 71 showed that manual correction of missed cones from an automated algorithm significantly increased repeatability, though newer automated algorithms have shown comparable repeatability. 68 Automatic cone identification becomes less reliable when rods begin scattering the mosaic or RPE cells become visible in eyes with retinal degeneration.…”