Abstract
For timely diagnosis of retinal disease, routine retinal monitoring of people with high risk should be put in place. To assist the ophthalmologists in performing retinal analysis efficiently and accurately, numerous studies have been conducted to propose an automated retinal diagnosis system. One of the crucial steps for such a system is accurate detection of retinal blood vessels from retinal image. In this paper, we investigated the use of automatic binarization methods on pre-processed fundus image to detect retinal blood vessels. Three methods for binarization were investigated in this study, namely Otsu’s method, ISODATA and K-means clustering method. The resulting binarized output indicated good detection of large vessels but most of the smaller vessels were left undetected. To address this issue, Gabor wavelet filter was used to enhance the small blood vessel structures before binarization of the filter output. Combining the binary images from both binarization with and without Gabor filter resulted in significant improvement of the overall detection rate of the retinal blood vessels. The proposed method proved to be comparable to other unsupervised techniques in the literature when validated using the publicly available fundus image database, DRIVE.
Highlights
Retinal images have been widely used for diagnosing multiple eye diseases through regular screening of patients’ retinal health
To cater for the increasing number of retinal images that needs to be diagnosed by the ophthalmologists on a daily basis, a lot of research have been conducted in these recent years in order to come up with a computer-assisted retinal diagnosis system
Sample images of post-processed binary images in Fig. 4 b) and d) indicate that the binarization of only the vessel-enhanced green channel image (GCI) may be sufficient in discovering most of the large vessels, but it can be noticed that most of the smaller vessels were left undetected
Summary
Retinal images have been widely used for diagnosing multiple eye diseases through regular screening of patients’ retinal health. Numerous underlying diseases such as diabetes mellitus and hypertension can be diagnosed using retinal image of the patient. To cater for the increasing number of retinal images that needs to be diagnosed by the ophthalmologists on a daily basis, a lot of research have been conducted in these recent years in order to come up with a computer-assisted retinal diagnosis system. One of the important pre-requisites for such a system is accurate segmentation of the retinal blood vessels from the input retinal image. That is why in these recent years large number of researchers have conducted studies into automatic segmentation of retinal blood vessels from retinal image. In [1] presented a comprehensive review of techniques used to detect blood vessels from retinal images in the recent years. The techniques were generally grouped into five categories including pattern recognition, matched filtering, morphological processing, vessel tracing, and multiscale methods
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal of Engineering & Technology
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
