China
Africa
Explore chapters and articles related to this topic
Imaging the Living Eye
Published in Margarida M. Barroso, Xavier Intes, In Vivo, 2020
Brian T. Soetikno, Lisa Beckmann, Hao F. Zhang
Other solutions for performing fundus photography have been reported. Paques et al. (2007) developed a low-cost, direct-contact solution by combining an illuminating endoscope with a digital camera. The eye is still illuminated with annular or crescent illumination to minimize specular corneal reflections. Standard laboratory microscopes can also be used to visualize the fundus by incorporating an additional objective lens or by applanation of the cornea with a contact lens or a microscope cover-slip (Hawes et al., 1999; Cohan et al., 2003). Finally, a handheld smartphone’s flash light and camera for illumination and observation, with the help of a standard indirect ophthalmoscope lens, was also demonstrated (Haddock et al., 2013).
Artificial Intelligence for Precision Medicine
Published in Kamal Kumar Sharma, Akhil Gupta, Bandana Sharma, Suman Lata Tripathi, Intelligent Communication and Automation Systems, 2021
Fundus photography, typically examined and assessed by ophthalmologists, is a non-invasive technique to capture images of the optic disc, the retina or the macula and discover or track eye diseases, for instance, age-related macular degeneration (AMD), glaucoma and diabetic retinopathy (DR). For instance, a DL algorithm was developed to find DR using 128 175 retinal images and reached analogous functioning to ophthalmologists in datasets of two independent examinations. Deep convolutional neural networks have been used to find glaucoma and instantly rank AMD using fundus photos, and the process reached a precision close to that of professional ophthalmologists [12].
Preliminary Study of Retinal Lesions Classification on Retinal Fundus Images for the Diagnosis of Retinal Diseases
Published in Mitul Kumar Ahirwal, Narendra D. Londhe, Anil Kumar, Artificial Intelligence Applications for Health Care, 2022
Jaskirat Kaur, Ramanpreet Kaur, Deepti Mittal
Fundus fluorescein angiography (FFA) is a procedure to image retina using a fundus camera by injecting a fluorescent dye. Sodium fluorescein dye is injected into the systemic circulation following which the retina is lit with blue light at a wavelength of 490 nm. Final FFA image is obtained by snapping the fluorescent green light that is emitted by the dye. FFA produces high contrast images of retina and is mainly used to analyse pathologies related to choroid and blood circulation that reside in retina. However, FFA being an invasive technique has a wide range of complications. The most common reactions are transient nausea, vomiting, pain and redness in eyes after the procedure.
Intelligent identification and classification of diabetic retinopathy using fuzzy inference system
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
Jyoti Prakash Medhi, R. Sandeep, Pranami Datta, Tousif Khan Nizami
These depositions over the retina block the vision and eventually lead to blindness. Early diagnosis is the only way out to prevent this irreversible vision loss. For retina analysis, ophthalmologists perform imaging using Fundus and Optical Coherence Tomography (OCT) cameras. Depending on the current patient-to-ophthalmologist ratio, this is a time-consuming procedure, vulnerable to human errors and devoid of reasonably good accuracy. This demands for a strong requirement of an automated screening mechanism, so that the ophthalmologists can judiciously concentrate more on treating the infected patients without loss of time and effort as mentioned by, Abràmoff et al. (2010); Medhi and Dandapat (2016a); Saeed and Oleszczuk (2016); Gharaibeh et al (2018, 2021). In view of the same, authors have proposed a novel supervised machine learning (ML) methodology to screen the DR and to further classify into its constituent stages.
A new approach to non-mydriatic portable fundus imaging
Published in Expert Review of Medical Devices, 2022
Faizal Hafiz, Renoh Johnson Chalakkal, Sheng Chiong Hong, Glenn Linde, Roger Hu, Ben O’Keeffe, Yim Boobin
Fundus imaging is a key to the effective diagnosis and treatment of ocular pathologies. Opportunistic diagnosis of such pathologies is often not sufficient as the treatment is critically dependent on early detection and diagnosis [1–3]. Given the increasing prevalence of diabetes mellitus, there is a need for a systematic large-scale screening program for diabetic retinopathy. For instance, national screening programs in Iceland and the UK have led to a significant reduction in blindness arising from diabetic retinopathy [1,2]. The replication of similar success in resource constrained environments is, however, challenging, owing to the immobility of stationary clinical settings associated with most state-of-the-art fundus imaging equipment [2,3]. While fundus imaging has matured over the years, it is mainly confined to stationary clinical settings due to the size and bulk of the associated equipment. A portable solution to fundus imaging is, therefore, crucial to increase outreach in under-served regions as well as to improve the efficacy under ‘telemedicine’ scenarios. This investigation, therefore, aims to develop a portable and cost-effective yet accurate fundus camera.
Automated segmentation of blood vessels in retinal images based on entropy weighted thresholding
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
Deepak Kumar Maharana, Pranati Das, Ranjeet Kumar Rout
Fundus images are colour photographs of the retina acquired by a fundus camera. A fundus camera is a low-power advanced microscope with an attached camera configured to image the eye’s inner surface, like the retina, optic disc, macula, etc. Ophthalmologists can diagnose numerous diseases, like diabetic retinopathy (DR) (Yau et al. 2012), blood pressure (Leung et al. 2004), cardiovascular disease (Wang et al. 2015), retinopathy of prematurity (Wilson et al. 2008), etc. Diabetic Retinopathy (DR) is caused by diabetes. The microvascular structures in the retinal image change due to diabetes. If the sugar level is high in diabetic patients, the tiny blood vessels throughout the retina may swell, rupture, or leak blood or other fluids on to the retinal surface (Frank 1995). Occasionally, the abnormal growth of new blood vessels on the retinal surface may cause impaired vision. Diabetic retinopathy (DR) can be separated into two types: proliferative diabetic retinopathy (PDR) and non-proliferative diabetic retinopathy (NPDR) (Paranjape and Kakatkar 2014). NPDR is the initial phase of diabetic retinopathy and can also be categorised as mild, moderate, or severe NPDR. If only one microaneurysm (MAs) is present in the retinal surface, then it is designated as mild NPDR (Basha and Prasad 2008). If more than one microaneurysm (MAs), haemorrhages, and exudates are present then it is termed as moderate NPDR (Usman Akram et al. 2014). The presence of more microaneurysms (MAs) and haemorrhages on the retinal surface may cause obstruction of the flow of blood in the vessels of the eye. Due to obstruction of the flow of blood in vessels, diabetic patients are affected by blurred vision. This is termed “Severe NPDR”. The advanced stage of DR is proliferative diabetic retinopathy (PDR) (Welikala et al. 2014). In the case of PDR, new abnormal blood vessels develop in the retinal region. In such a case, it is a very tough task to discriminate between normal and abnormal vessels. The thorough assessment of DR is shown in Table 1 (Chalakkal et al. 2020). So, to accurately visualise and quantify the pathological disorders in the retina, vasculature segmentation is the basic step.