Computational efficiency of optic disk detection on fundus image: a survey

Abstract

Fundus image processing is getting widely used in retinopathy detection. Detection approaches always proceed to identify the retinal components, where optic disk is one of the principal ones. It is characterized by: a higher brightness compared to the eye fundus, a circular shape and convergence of blood vessels on it. As a consequence, different approaches for optic disk detection have been proposed. To ensure a higher performing detection, those approaches varied in terms of characteristics set chosen to detect the optic disk. Even the performances are slightly different, we distinguish a significant gap on the computational complexity and hence on the execution time. This paper focuses on the survey of the approaches for optic disk detection. To identify an efficient approach, it is relevant to explore the chosen characteristics and the proposed processing to locate the optic disk. For this purpose, we analyze the computational complexity of each detection approach. Then, we propose a classification approach in terms of computational efficiency. In this comparison study, we distinguish a relation between computational complexity and the characteristic set for OD detection.

Retinal blood vessel segmentation from high resolution fundus image using deep learning architecture

“…This approach offers the possibility of applying the proposed technique across various datasets, in order to perform an accurate feature extraction. The same DL extension can be adopted for the segmentation of artery and vein, which is crucial to detect several retinal components [12] or pathologies such hypertensive retinopathy [13], cataract [14], diabetic retinopathy [15] and aged macular degeneration [16].…”

A Fast and Accurate Method for Glaucoma Screening from Smartphone-Captured Fundus Images

“…The method proposed in [44] ensured locating the ONH based on circularity and intensity through applying the radon transform. The method achieved higher accurate performance among the existing methods [31,34], where 100% accuracy was performed to locate the ONH in the public DRIVE database. We note that this method can be configured to locate a sub-image having the same OD size.…”

Efficient graphical-processor-unit parallelization algorithm for computing Eigen values

“…Moreover, several methods proceed to segment separately thick and thin vessel tree and thereafter merge their results in order to enhance accuracy when segmenting the whole tree of retinal vessel 41,42 . Furthermore, the segmentation of thick vessel tree is considered as main step on automated methods for retinal component detection, where thick vessel convergence to the optic disk 43 and leak of thick vessel on macula region 44 are widely used to locate them. Particularly, the thick vessel tree is segmented in 3 based on Eigen Values, where the result was depicted in Fig.…”