Glaucoma detection

Archana et al. [14] developed a novel approach for glaucoma detection. The main motive of this study is to map the person’s eye color image with database of images consisting of normal person as well as glaucoma affected person. The different color variations inside the eye can be compared by using images taken from high definition laser camera. They are also known as fundus images. The feature extraction of these fundus images may be carried out using MATLAB software tool. By measuring the color pixels in the affected area the observation shows that the person is suffering from Glaucoma or not. To detect if a person is suffering from Glaucoma or not, a test is conducted using the image of a normal person which is kept as reference and then several comparison are carried out with the clinical observations of the person’s image. Further if the result is positive (person is affected with Glaucoma) then also check for the three types of Glaucoma i.e. Primary Angle Closure Glaucoma, Secondary Glaucoma, and Congenital Glaucoma.

Optic Nerve Head (ONH) assessment for glaucoma detection is commonly done by observing stereo optic disc images for structural abnormalities. Stereoscopic photographs of the optic nerve allow a perception of the cup depth, which is an important cue for glaucoma. There are many researches in monocular images that have evaluated the changes in the optic nerve by computing the ONH depth map using stereo matching methods [7-8]. Cup-to-Disc Ratio (CDR) is an important parameter which is calculated from the optic disc and cup and is commonly used in identifying elongation of optic cup and loss of neuron-retinal rim. In [9], CDR estimation with thresholding techniques is calculated from stereo images. This method requires the huge cost and high

ABSTRACT: This paper presents automated glaucoma detection techniques based on neural network and Adaptive Neuro fuzzy Inference system (AN- FIS) Classifier. Digital image processing techniques, such as preprocessing, morphological operations and thresholding, are widely used for the auto- matic detection of optic disc, blood vessels and computation of the features of fundus image. Glaucoma is a disease of the optic nerve caused by the increase in the intraocular pressure of the eye. Glaucoma mainly affects the optic disc by increasing the cup size. It can lead to the blindness if it is not detected and treated in proper time. The detection of glaucoma through Optical Coherence Tomography (OCT) and Heidelberg Retinal Tomography (HRT) is very expensive, this limitation is removed by this Glaucoma Diagnosis system with good performance. In addition to diagnosis of Glaucoma a Graphical user interface (GUI) is developed. This GUI is used for automatic diagnosing and displaying the diagnosis result in a more friendly user inter- face The results presented in this paper indicate that the features are clinically significant in the detection of glaucoma. Proposed system of this paper is able to classify the glaucoma automatically with a sensitivity and specificity of 98% and 95% respectively.

try in the ganglion cell layer showed the best diagnostic results for early glaucoma detection. Most these previous studies used the asymmetry of macular thickness rather than the pRN- FL thickness, because there is less variability in the former. However, the use of asymmetry in macular thickness also has its limitations. First, the use of macular thickness asymmetry is limited when there is concurrent macular disease. As glau- coma is common in older patients, many glaucoma patients have concurrent macular disease. Second, damage outside the macular area may not be detected by macular analysis. How- ever, to the best of our knowledge, only one study has reported so far on the diagnostic ability of asymmetry in pRNFL thick- ness between the superior and inferior areas. 18 In contrary to

ABSTRACT:There are several papers are available on different types for glaucoma detection. Here we are trying to understand possible some techniques and try to represent compensative approach on different detection techniques such as Funduscopic Images, Angio- Optical Coherence Tomography (OCT) Images, Retinal Fundus Images, Colour Fundus Images (CFI).Paper based on compensative review of different glaucoma detection techniques is presented in this paper. The paper also discusses various structural features that are related to CFI and OCT images respectively for automated glaucoma detection.

The proposed system will be very useful to detect glaucoma efficiently so that the disease can be diagnosed in the early stages. The system does not depend on trained glaucoma specialist and expensive HRT/OCT machines. In this project the glaucoma detection is done by extracting three features i.e. optic disk, optic cup and neuro-retinal rim from a digital fundus image. The artificial neural networks is used as a classifier to identify the disease. The thresholding approach is used for optic disk and optic cup segmentation.

ABSTRACT: The aim of this paper is to design and implement an automated system for glaucoma detection using cup to disc ratio. Glaucoma is a chronic eye disease affecting the optic nerve which leads to blindness. It is second leading cause for blindness. Glaucoma affects Optic nerve head. Some methods for glaucoma detection include measurement of intraocular pressure, assessment of abnormal visual field and assessment of damaged optic nerve head. Calculation of cup to disc ratio is one of the methods in assessment of damaged optic nerve head. Optic Disc and Optic Cup are parts of nerve head. First, retinal fundus image is acquired. Then optic disc is localized using thresholding. This is preprocessing step. Accurate localization is very important marker for many computer aided design techniques. Superpixel segmentation (which uses simple linear iterative clustering algorithm) is used to extract correct disc and cup boundaries. Number of superpixels is the only parameter for superpixel segmentation. Center Surround statistics is used for feature extraction purpose. This is because area around disc looks similar to disc(Paripappilary Atrophy) in color but differs in texture. Artificial neural network is used as classifier. Then vertical cup diameter to vertical disc diameter ratio is calculated. If it is greater than normal value then that is glaucomatous. Objectives of project are correct localization of optic disc and finding cup inside disc. Resulting images show disc and cup boundaries and cup to disc ratio. Sensitivity and accuracy are two main parameters for evaluation purpose. Success rate achieved in this method is 89.18%.

There were 14 patients (10 men, 4 women) enrolled in this study, ranging in age from 51 to 80 years (mean age, 61.4 years). The patients comprised cases of normal-ten- sion glaucoma (NTG, 6 eyes) and primary open-angle glaucoma (POAG, 8 eyes). The inclusion criteria required a corrected visual acuity of 1.0 (= 0 logMAR) or better and a pupil size of at least 2.5 mm without dilation. Among the exclusion criteria of the patients were severe cataracts (grade III to V in the Emery-Little classification) and drugs affecting the pupil, particularly pilocarpine. Also, this study examined patients without any systemic or oph- thalmic diseases likely to affect the visual field (apart from glaucoma). Prior to the study, all patients were examined with a Humphrey Field Analyzer (30-2, Full threshold program). The mean deviation (MD) value ranged from - 6.50 dB to - 20.18 dB. Patients with more than 20% false-positive or false-negative responses were excluded. A diagnosis of POAG was based on Anderson and Patella’s criteria, 13 as well as on a glaucomatous optic disc, and an

With great improvement in field of medical imaging, Image processing technique helps in early diagnosis of glaucoma and other eye disease. Retinal fundus images assist trained clinicians to diagnose any abnormality and any change in retina. These images are captured by using special devices called ophthalmoscopes. Medical image analysis and processing has great significance in non-invasive treatment and clinical study. The information about the optic disk can be used to examine severity glaucoma. The location of the optic disk is an important issue in retinal image analysis as it is a significant landmark feature. Fig.1 shows the fundus camera and retinal fundus image.

Glaucoma is disease related to optic nerve that arises due to deviation in fluid density inside the eye in addition harms the optic nerve [1, 2].Glaucoma, an optic nerve disease occurs due to variation – increase or decrease in fluid pressure within the eye. The pressure of the normal eye is 21 mm of Hg and when the pressure value is higher than 21 mmHg or 2.8 kPa, the optic nerve gets damaged leading to vision impairments and blindness. Diagnosis and treatment is based on elevation of Intraocular Pressure (IOP). Due to the buildup pressure in the eye, the nerve cells become compressed eventually leading to permanent vision loss. The anterior chamber of eye is bounded by cornea, iris, pupil and lens which is filled by fluid called aqueous humor. Aqueous humor nourishes the lens and cornea with nutrients, oxygen and provides optimal pressure called as IOP to maintain the shape of the eye. IOP is measured for detecting glaucoma and diagnosis is done by dilated eye examination showing an abnormal amount of cupping [3]. The post symptoms of Glaucoma are seeing halos around lights, vision loss, Redness in the eye, nausea or vomiting, eye pain and narrowed vision (tunnel vision) [4].

M. R. K. Mookiah, U. R. Acharya, C. M. Lim, A. Petznick, and J. S.Suri,et al., developed a new method for the automated identification of normal and glaucoma classes using Higher Order Spectra (HOS) and Discrete Wavelet Transform (DWT) features (M. R. K. Mookiah et al, 2013). The extracted features are fed to the Support Vector Machine (SVM) classifier with linear, polynomial order 1, 2, 3 and Radial Basis Function (RBF) to select the best kernel function for automated decision making. HOS consist of moment and cumulant spectra. It can be used for both deterministic signals and random. Higher order spectra invariants have been used for shape recognition and to identify different kinds of eye disease. HOS is a non- linear method helps to capture the subtle changes in pixels of the image, which can be used as features for the automated classification. A given signal is decomposed using DWT into approximation and detail coefficients at first level. The approximation coefficients are further decomposed to obtain the next level approximation and detail coefficients to provide DWT representation at a higher level of decomposition. In the first step HOS and DWT features extracted from the digital fundus images. Support Vector Machine (SVM) is a supervised learning methods used for classification and regression. Extracted feature are used for classification of normal and Glaucoma images The higher values of HOS features and lower values of wavelet energy features suggest that glaucoma has a more coarse textural variation than normal.

A random selection of 100 optometrists was recruited as detailed in a previous publication.(18) Although the optometrists were expecting visits from SPs, steps were taken to ensure that the SP remained undetected. Participating optometrists were not given any information about the characteristics of the SP, nor were they aware that their ability to investigate appropriately a patient in a high risk group for glaucoma was being assessed. No SP visits took place within a month of the optometrist recruitment. The SP is a professional actor and, prior to visiting consenting optometrists, she underwent intensive one-to-one training on the different aspects of an eye examination.(19) This involved use of a document entitled “The journey through an eye examination” which describes an eye examination in lay terms. The actor observed and received several eye examinations (some whilst being observed) from different clinicians at the Institute of Optometry. The actor was trained to remember and record details of the clinical encounter. Some eye examinations during the training were video recorded to allow for quality control later in the study when it was felt that it would be helpful to remind the SP of certain tests. The SP was also given a copy of a video of one of their training eye examinations on a CD. At the end of the training the actor signed a confidentiality agreement stating that any information gathered during the eye examinations is confidential and will be used solely for the completion of the checklist provided.

Feed forward neural network (FFN) is used to detect diseases like hypertension and glaucoma.Detection of glaucoma, hypertension are based on feature extraction. So Gray Level Co-occurrence Matrix is used. It calculates pairs of pixel with specific values and in a specified spatial relationship occur in a image. For feature extraction, the features of GLCM are Contrast, Correlation, Energy and Homogeneity.Additionally the other features are Mean and Standard deviation of Red, Green , Blue ,Hue, Saturation and Intensities in the vessel. A FFN is an artificial neural network where information moves only in one direction i.e. forward from input node .There are hidden nodes in between input nodes and output nodes. The reason for hypertension is due to high blood pressure. High blood pressure can damage the vessels in retina. Their Symptoms are Double vision, vision loss, Headaches. Glaucoma is a type of eye disorder that creates damage in optic nerve. It isassociated with increased fluid pressure in the eye known as intraocular pressure.

In the recent years, the number of persons identified with diabetic retinopathy has increased and the glaucoma detection also posed many challenging tasks. Glaucoma is an eye disorder that affects the optic nerve to have permanent vision loss. In general, the glaucoma can be detected through various methods by analyzing physiological parameters of the eye. The image processing technique using fundus images, Ultrasound images and optic disc photographs have been utilized by many researchers to identify glaucoma and hence the diabetic retinopathy. The major symptoms of glaucoma include (1) Blurred vision (2) Severe pain in the eye (3) Rainbow hallows with light Headache (4) Brow pain Nausea (5) Vomiting with Red Eye. The intraocular pressure is also identified as one of the risk factors which develop the glaucomatous damage and lower pressure leads to progressive retinal degenerative change.

and wavelet transforms are applied on fundus image to extract high frequency components for cataract classification. It consists of fundus image pre-processing, feature extraction, followed by cataract classification and grading. S Dua [8] used Wavelet-Based Energy Features to compare the performance of different classifiers in glaucoma detection: Naïve Based Classifier, k-Nearest Neighbor Classifier and Support Vector Machines (SVM), Random Forests and Sequential minimal optimization (SMO). Here energy features are extracted from retinal images using five different wavelets (Daubechies, Symlet Bior 3.3, Bior 3.5 and Bior 3.7). They observed minimum classification efficiency from SMO.

OD segmentation method is proposed which integrates the local image information around each point of interest in multidimensional feature space to provide robustness against variations found in and around the OD region. A new template-based methodology for segmenting the OD from digital retinal images is presented in [9]. Morphological and edge detection techniques followed by the Circular Hough Transform are used to obtain a circular OD.Within the OD as initial information. A novel method for glaucoma detection using a combination of texture and higher order spectra (HOS) features from digital fundus images is proposed in [10]. Support vector machine, sequential minimal optimization, naive Bayesian, and random-forest classifiers are used to perform supervised classification. A mathematical framework to link retinal nerve fiber layer (RNFL) structure and visual function using the data typically acquired in the clinical management of glaucoma is proposed in [11]. The model performed and generalized well over different populations from three clinical centers. The derived structure-function relationship accorded well with RNFL anatomy, and could be applied to reduce the variability that confounds the measurement of glaucoma damage.Stereo disc photograph is analyzed and reconstructed as 3 dimensional contour images to evaluate the status of the optic nerve head for the early detection of glaucoma and the evaluation of the efficacy of treatment is presented in [12]. To detect the edge of the optic nerve head and retinal vessels and to reduce noises, stepwise preprocessing is introduced. RetCam is a new imaging modality that captures the image of iridocorneal angle for the classification is presented in [13]. Glaucoma is the one of the two major causes of blindness, which can be diagnosed through measurement of neuro-retinal CDR is described in [14]. Automatic calculation of optic cup boundary is challenging due to the interweavement of blood vessels with the surrounding tissues around the cup. A multimodality fusion approach for neuro retinal cup detection improves the accuracy of the boundary estimation. Modeling of Scanning Laser polarimetry method is presented in [15] to model the change in images acquired by scanning laser polarimetry for the detection of glaucomatous progression.. Sobel edge detection, Texture Analysis, Intensity and Template matching techniques are used to detect Optic Disc. A system has been proposed to detect glaucoma using fundus image [16] consisting three different stages has ROI extensions, feature extraction stage and classification stage.

Glaucoma is characterized by dysfunction and loss of retinal ganglion cells (RGCs), with resulting structural changes to the optic nerve head, retinal nerve fiber layer (RNFL) thickness, and ganglion cell and inner plexiform layers as well as loss of the visual field.The diagnosis of glaucoma in its early stages is challenging. Misdiagnosis can lead to failure to identify individuals with the condition during its early stages until significant functional loss has occurred. Thus, early detection of glaucoma allows for early treatment to delay vision loss. Diagnosing glaucoma is problematic, especially when it is in the earliest stage of glaucoma. Diagnosis of glaucoma in myopic eyes and patients with brain diseases such as brain tumor is known to be difficult due to those eye‟s characteristic disc shape and visual field defect. A more effective glaucoma – detection machine learning model would be very helpful to medical practitioners.The classification scheme in machine learning is suitable for diagnosis glaucoma. Various classification algorithms were tested. The classification algorithms used were theK – Nearest Neighbor, Decision Tree, Random Forest Classifier.

In order to overcome these difficulties, automated diagnosis methods are preferred for glaucoma diagnosis. Selections of robust features are necessary to develop a robust system. Recent studies have shown that texture features are very effective for glaucoma image detection. Higher Order Spectra (HOS) coupled with texture features are used to improve the classification accuracy. In, Discrete Wavelet Transform (DWT) energies are used as features for glaucoma detection. The HOS bi-spectrum features and wavelet energy features are used for glaucoma diagnosis in eye. The DWT has a set of fixed basis functions that are signal independent. The working principle of EWT depends upon frequency spectrum of the signal. In this work, we are proposing a novel method for the classification of glaucoma images based on EWT and correntropy features. EWT decomposes the image into various frequency bands. Correntropy is extracted from the decomposed EWT components. The features are normalized and ranked on the basis of significant criteria. Least Squares Support Vector Machine (LS-SVM) classifier with various kernels such as Radial Basis Function (RBF) and wavelet kernels such as Morlet and Mexican-hat are used for the classification. Three-fold and ten-fold cross validation strategies are used to develop the automated glaucoma diagnosis system.

The laboratory experimental results are obtained in order to detect neovascular glaucoma results are tabulated in Table1.Table1 reveals that fractal dimension values which ranges from 0.98 to 1.61 for healthy images and for Neo- vascular glaucoma exceeds the upper limits and maximum 1.99 for retinal images for different threshold values of binary images. The results of the proposed research work emphasizes that there is great potential in the use of fractal dimension for estimating non glaucoma and Neo-vascular glaucoma retinal images. The proposed approach achieved better results compared to the results reported in literature. An extension for this research work could be as follows: (1) Instead of consider work retinal image can be considering only neovasculization portion of retina.

This research was a prospective clinical study carried out in several rural communities of southeast Nigeria between December 2010 and February 2011. Subjects were assembled from neighboring rural communities in two states of the southeast geographical region on Nigeria. An informed consent was obtained from all the subjects used for this study. Examination of subjects involved taking of case history, external examination using pen light, visual acuity test using the Snellen visual acuity chart, retinoscopy with the Keeler retinoscope, ophthalmoscopy with the Keeler ophthalmoscope, indentation tonometry using schoitz tonometer, subjective refraction and visual field test. The blood pressure was taken with the use of KODEA electronic sphygmomanometer. The subjects who had both glaucoma and hypertension were then noted. RESULTS