Normal Eye Images Research Articles

Design of A Cataract Detection System based on The Convolutional Neural Network

Cataract, a condition characterized by clouding of the eye's lens, leads to decreased vision and potentially blindness. In Indonesia, it is the predominant cause of blindness, accounting for 81.2% of cases. Given the rising life expectancy, the incidence of degenerative diseases like cataracts is expected to increase. This research aims to develop a cataract detection system capable of classifying eye images as either indicative of cataracts or normal. Utilizing Convolutional Neural Networks (CNN) and RGB-based image processing—including edge detection techniques such as Canny and Prewitt—the system identifies eye contours. This facilitates image segmentation to ascertain the eye's condition. Therefore, image collection and processing models play a crucial role in this study. The research findings indicate that the system functions effectively, with a 98% success rate in accurately processing normal eye images through the CNN model without detecting cataracts. When tested using grayscale imaging, cataract-afflicted eyes—characterized by red spots in the images—were also successfully identified by the CNN model. These test results demonstrate that the designed cataract detection system can accurately classify images into normal or cataract-afflicted eyes with high precision. This system shows promise for use in early cataract detection, potentially helping to reduce the prevalence of cataract-related blindness in Indonesia.

Drugged Eye Detection Using Image Processing

Abstract: Drugs are a major problem in economic and many losses in worldwide. In this project, an image processing approach is proposed for identifying drugged eye based on convolutional neural network. According to the CNN algorithm, eye image details are taken by the existing packages from the front end used in this project. However, it can take a few moments. So, this proposed system can be used to identify drugged eyes quickly and automatically. The eye images dataset are taken from Kaggle. These images are taken as a training set for this drugged eye detection. This proposed approach is composed of the following main steps that getting input image, Image Preprocessing, identifying reddish places, highlight those affected places, Verifying training set, showing result. Few types of eyes like drugged socially may missed to identify. This approach was tested according to drugged eye type and its' stages, such as drug consumed and not consumed. The algorithm was used for detecting the white area of eye present in given input image. Images were provided for training, such as drugged eye images and normal eye images. Before the image processing, images were converted to color models, because of find out the most suitable color model for this approach. Local Binary Pattern was used for feature extraction and Support erosion method was used for creating the model. According to this approach, drugged eyes can be identified in the average accuracy of 95%.

Denoising and contrast enhancement of normal eye images and slit lamp images of cataract using optimized deep learning model

Denoising and contrast enhancement techniques in normal and slit lamp images are challenging in medical analysis. These images are essential in medical analysis for preventing, diagnosing, and monitoring various diseases. For this purpose, the medical images should be clear. Hence, this work proposes an optimized deep learning model to provide denoised and contrast-enhanced images. To begin, the selection of picture quality is merged to identify high-quality normal as well as slit lamp images for subsequent analysis. The picture quality of an image is examined with the model “Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE)”. It yields a low score for pictures of high-quality, an average score for pictures of medium quality, as well as a high score for images of poor quality. Furthermore, normalization-based contrast limited adaptive histogram equalisation (N-CLAHE) is a new algorithm for converting medium- and low-quality images to high-quality images. The images are then pre-processed to remove noise using wiener filtering (WF) with a convolutional neural network (CNN) with adaptive atom search optimization (CNN-AASO). Further, the denoised image is enhanced by Gaussian mixture-based contrast enhancement (GMCE) for contrast enhancement. The overall implementation is carried out on the Matlab tool.

A Detection of Amblyopia Medical Condition in Biomedical Datasets Using Image Segmentation and Detection Processing

The recent past, the data volume in a media field is growing at a rapid rate, and conventional methods fail to manage such a large volume of data in healthcare systems, biomedical field, medical diagnostic systems etc. The main challenges associated with biomedical computation are the problems associated with management, storage, and analysis on extensive biomedical data. To play a significant role over such extensive data, the machine learning approach provides faster access to medical data with an improved framework. The main objective involves the detection of amblyopia condition from input images and comparing it with conventional image detection methods. The proposed method is examined in terms of detection accuracy, sensitivity, specificity, Hausdorff distance computation and Dice Coefficient. Also, the detection of an Amblyopic or Lazy Eye diseased images is still not prevalent in the field of image segmentation and detection. In this paper, we introduce a framework to process the Amblyopia image datasets using machine learning, and similarity comparison approach. The proposed image processing involves the segmentation of eye images using Recurrent Neural Networks (RNN), and the detection of Amblyopia disease is carried out with Hausdorff Distance computation and Dice coefficient similarity comparison on the segmented image. The initial subset points and threshold values are calculated from a set of 50 normal eye images. A set of 100 Amblyopic diseased image dataset is used for testing the proposed system, out of which 70 images are used for training the system. To evaluate the experimental results shows that proposed method obtains improved detection than existing Deeply-Learned Gaze Shifting Path (DLGSP), Cascade Regression Framework (CRF) and Mobile Iris Recognition System (MIRS) methods. The presence of Hausdorff Distance computation and Dice coefficient similarity comparison is used for reducing the overhead in the proposed method, and this can be used for computing large sets of images.

Application of convolution neural network in medical image processing.

Convolution neural network is often superior to other similar algorithms in image classification. Convolution layer and sub-sampling layer have the function of extracting sample features, and the feature of sharing weights greatly reduces the training parameters of the network. This paper describes the improved convolution neural network structure, including convolution layer, sub-sampling layer and full connection layer. This paper also introduces five kinds of diseases and normal eye images reflected by the blood filament of the eyeball "yan.mat" data set, convenient to use MATLAB software for calculation. In this paper, we improve the structure of the classical LeNet-5 convolutional neural network, and design a network structure with different convolution kernels, different sub-sampling methods and different classifiers, and use this structure to solve the problem of ocular bloodstream disease recognition. The experimental results show that the improved convolutional neural network structure is ideal for the recognition of eye blood silk data set, which shows that the convolution neural network has the characteristics of strong classification and strong robustness. The improved structure can classify the diseases reflected by eyeball bloodstain well.

Optimasi decision tree menggunakan particle swarm optimization untuk identifikasi penyakit mata berdasarkan analisis tekstur

The problem of visual impairment is a serious problem with increasing cases, ranging from visual impairment to the cause of blindness. This study examines the development of an identification application for the classification of patients with eye disorders using the Decision Tree (DT) method, which is optimized using Particle Swarm Optimization (PSO). This study used 311 eye image data, consisting of 233 normal eye images and 78 eye images with glaucoma, cataracts, and uveitis. The feature extraction used Gray Level Co-occurrence Matrix (GLCM), while the feature optimization used the PSO and the learning method used DT. This optimized visual impairment classification application can improve system accuracy to 88.09 %.

Deep learning-based automated detection of glaucomatous optic neuropathy on color fundus photographs.

To develop a deep learning approach based on deep residual neural network (ResNet101) for the automated detection of glaucomatous optic neuropathy (GON) using color fundus images, understand the process by which the model makes predictions, and explore the effect of the integration of fundus images and the medical history data from patients. A total of 34,279 fundus images and the corresponding medical history data were retrospectively collected from cohorts of 2371 adult patients, and these images were labeled by 8 glaucoma experts, in which 26,585 fundus images (12,618 images with GON-confirmed eyes, 1114 images with GON-suspected eyes, and 12,853 NORMAL eye images) were included. We adopted 10-fold cross-validation strategy to train and optimize our model. This model was tested in an independent testing dataset consisting of 3481 images (1524 images from NORMAL eyes, 1442 images from GON-confirmed eyes, and 515 images from GON-suspected eyes) from 249 patients. Moreover, the performance of the best model was compared with results obtained by two experts. Accuracy, sensitivity, specificity, kappa value, and area under receiver operating characteristic (AUC) were calculated. Further, we performed qualitative evaluation of model predictions and occlusion testing. Finally, we assessed the effect of integrating medical history data in the final classification. In a multiclass comparison between GON-confirmed eyes, GON-suspected eyes and NORMAL eyes, our model achieved 0.941 (95% confidence interval [CI], 0.936-0.946) accuracy, 0.957 (95% CI, 0.953-0.961) sensitivity, and 0.929 (95% CI, 0.923-0.935) specificity. The AUC distinguishing referrals (GON-confirmed and GON-suspected eyes) from observation was 0.992 (95% CI, 0.991-0.993). Our best model had a kappa value of 0.927, while the two experts' kappa values were 0.928 and 0.925 independently. The best 2 binary classifiers distinguishing GON-confirmed/GON-suspected eyes from NORMAL eyes obtained 0.955, 0.965 accuracy, 0.977, 0.998 sensitivity, and 0.929, 0.954 specificity, while the AUC was 0.992, 0.999 respectively. Additionally, the occlusion testing showed that our model identified the neuroretinal rim region, retinal nerve fiber layer (RNFL) defect areas (superior or inferior) as the most important parts for the discrimination of GON, which evaluated fundus images in a way similar to clinicians. Finally, the results of integration of fundus images with medical history data showed a slight improvement in sensitivity and specificity with similar AUCs. This approach could discriminate GON with high accuracy, sensitivity, specificity, and AUC using color fundus photographs. It may provide a second opinion on the diagnosis of glaucoma to the specialist quickly, efficiently and at low cost, and assist doctors and the public in large-scale screening for glaucoma.

Detection of glaucoma using two dimensional tensor empirical wavelet transform

Retinal image analysis is extensively used in the medical field for the detection of abnormalities in the eye. Amongst the several abnormalities in the eye, glaucoma has been identified as the second major cause of blindness. The glaucoma disease are of various types, out of which close angle glaucoma causes more pain and has the most effect on eye vision in comparison with the open angle glaucoma. In this paper, a futuristic computer aided classification method is designed and differentiated in terms of its sensitivity, specificity, and classification accuracy. The proposed method is used to detect glaucoma eye fundus from normal eye images using tensor based empirical wavelet transform (tensor based EWT). The method was tested on samples obtained from the online database; Medical image group Analysis, and an accuracy of 98.7% detection rate was achieved. The tensor based EWT yields 22% improvement in comparison with the littlewood EWT. The performance of the Tensor based EWT technique was compared with existing techniques and was found to outperform existing results.

Image processing-based model for tortuosity measurement of retinal blood vessels

Purpose: The study of retinal blood vessel morphology is of great importance in retinal image analysis. The retinal blood vessels have a number of distinct features such as width, diameter, tortuosity, etc. In this paper, a method is proposed to measure the tortuosity of retinal blood vessels obtained from retinal fundus images. Tortuosity is a situation in which blood vessels become tortuous, that is, curved or non-smooth. It is one of the earliest changes that occur in blood vessels in some retinal diseases. Hence, its detection at an early stage can prevent the progression of retinal diseases such as diabetic retinopathy, hypertensive retinopathy, retinopathy of prematurity, etc. The present study focuses on the measurement of retinal blood vessel tortuosity for the analysis of hypertensive retinopathy. Hypertensive retinopathy is a condition in which the retinal vessels undergo changes and become tortuous due to long term high blood pressure. Early recognition of hypertensive retinopathy signs remains an important step in determining the target-organ damage and risk assessment of hypertensive patients. Hence, this paper attempts to estimate tortuosity using image-processing techniques that have been tested on artery and vein segments of retinal images.
 Design: Image processing-based model designed to measure blood vessel tortuosity.
 Methods: In this paper, a novel image processing-based model is proposed for tortuosity measurement. This parameter will be helpful for analyzing hypertensive retinopathy. To test the eff ectiveness of the system in determining tortuosity, the method is first applied on a set of synthetically generated blood vessels. Then, the method is repeated on blood vessel (both artery and vein) segments extracted from retinal images collected from publicly available databases and on images collected from a local eye hospital. The blood vessel segment images that are used in the method are binary images where blood vessels are represented by white pixels (foreground), while black pixels represent the background. Vessels are then classified into normal, moderately tortuous, and severely tortuous by following the analysis performed on the images in the Retinal Vessel Tortuosity Data Set (RET-TORT) obtained from BioIm Lab, Laboratory of Biomedical Imaging (Padova, Italy). This database consists of 30 artery segments and 30 vein segments, which were manually ordered on the basis of increasing tortuosity by Dr. S. Piermarocchi, a retinal specialist belonging to the Department of Ophthalmology of the University of Padova (Italy). The artery and vein segments with the fewest number of turns are given a low tortuosity ranking, while those with the greatest number of turns are given a high tortuosity ranking by the expert. Based on this concept, our proposed method defines retinal image segments as normal when they present the fewest number of twists/turns, moderately tortuous when they present more twists/turns than normal but fewer than severely tortuous vessels, and severely tortuous when they present a greater number of twists/turns than moderately tortuous vessels. On implementing our image processing-based method on binary blood vessel segment images that are represented by white pixels, it is found that the vessel pixel (white pixels) count increases with increasing vessel tortuosity. That is, for normal blood vessels, the white pixel count is less compared to moderately tortuous and severely tortuous vessels. It should be noted that the images obtained from the different databases and from the local hospital for this experiment are not hypertensive retinopathy images. Instead, they are mostly normal eye images and very few of them show tortuous blood vessels.
 Results: The results of the synthetically generated vessel segment images from the baseline for the evaluation of retinal blood vessel tortuosity. The proposed method is then applied on the retinal vessel segments that are obtained from the DRIVE and HRF databases, respectively. Finally, to evaluate the capability of the proposed method in determining the tortuosity level of the blood vessels, the method is tested with a standard tortuous database, namely, the RET-TORT database. The results are then compared with the tortuosity level mentioned in the database. It was found that our method is able to classify blood vessel images as normal, moderately tortuous, and severely tortuous, with results closely matching the clinical ordering provided by the expert in the RET-TORT database. Subjective evaluation was also performed by research scholars and postgraduate students to cross-validate the results.
 Conclusion: The close correlation between the tortuosity evaluation using the proposed method and the clinical ordering of retinal vessels as provided by the retinal specialist in the RET-TORT database shows that, although simple, this method can evaluate the tortuosity of vessel segments effectively.

Eye Disease in Patients with Diabetes Screened with Telemedicine.

Telemedicine with nonmydriatic cameras can detect not only diabetic retinopathy but also other eye disease. To determine the prevalence of eye diseases detected by telemedicine in a population with a high prevalence of minority and American Indian/Alaskan Native (AI/AN) ethnicities. We recruited diabetic patients 18 years and older and used telemedicine with nonmydriatic cameras to detect eye disease. Two trained readers graded the images for diabetic retinopathy, age-related macular degeneration (ARMD), glaucomatous features, macular edema, and other eye disease using a standard protocol. We included both eyes for analysis and excluded images that were too poor to grade. We included 820 eyes from 424 patients with 72.3% nonwhite ethnicity and 50.3% AI/AN heritage. While 283/424 (66.7%) patients had normal eye images, 120/424 (28.3%) had one disease identified; 15/424 (3.5%) had two diseases; and 6/424 (1.4%) had three diseases in one or both eyes. After diabetic retinopathy (104/424, 24.5%), the most common eye diseases were glaucomatous features (44/424, 10.4%) and dry ARMD (24/424, 5.7%). Seventeen percent (72/424, 17.0%) showed eye disease other than diabetic retinopathy. Telemedicine with nonmydriatic cameras detected diabetic retinopathy, as well as other visually significant eye disease. This suggests that a diabetic retinopathy screening program needs to detect and report other eye disease, including glaucoma and macular disease.

The Emotional Eye: Red Sclera as a Uniquely Human Cue of Emotion

AbstractThe sclera, the eye's tough white outer layer, provides the ground necessary for the display of its own color and that of the overlying membrane, the conjunctiva. This study evaluated the sclera as a cue of emotion by contrasting the ratings of 38 subjects for the level of anger, fear, sadness, disgust, happiness, or surprise of normal (untinted) eye images with copies of those images that were reddened by digital editing. Subjects rated individuals with reddened sclera as having more anger, fear, disgust, and sadness, and less happiness than those with normal, untinted sclera. Surprise was the only emotion unaffected by scleral redness. Humans, but not other primates, have evolved the white sclera necessary to display the blood flow in the overlying conjunctiva that produces the redness associated with certain emotional states.

Changes in the Amygdala Produced by Viewing Strabismic Eyes

The aim of this study was to look for the response to strabismus images in the limbic network (amygdala, hippocampus, parahippocampus) of healthy volunteers and to compare it with their reaction to viewing normal eyes. Prospective, observational study. Thirty-one healthy volunteers underwent functional magnetic resonance imaging (fMRI). Functional magnetic resonance imaging data and blood oxygen level-dependent signal changes were analyzed using the BrainVoyager QX software package (Brain Innovation, Maastricht, The Netherlands). Responses to viewing strabismus images were compared with those observed while viewing normal eye images. Strabismus images led to significant activation of the amygdala, hippocampus, parahippocampal, and fusiform gyri in 30 of 31 subjects compared with normal eye images, indicating a negative emotional response. These fMRI results confirm that strabismus influences organically not only the patient with nonparallel eyes but also observers. Treatment of strabismus therefore changes the interpersonal dynamic for patients with strabismus on a demonstrable organic basis. The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Automatic identification of diabetic maculopathy stages using fundus images

Diabetes mellitus is a major cause of visual impairment and blindness. Twenty years after the onset of diabetes, almost all patients with type 1 diabetes and over 60% of patients with type 2 diabetes will have some degree of retinopathy. Prolonged diabetes retinopathy leads to maculopathy, which impairs the normal vision depending on the severity of damage of the macula. This paper presents a computer-based intelligent system for the identification of clinically significant maculopathy, non-clinically significant maculopathy and normal fundus eye images. Features are extracted from these raw fundus images which are then fed to the classifier. Our protocol uses feed-forward architecture in an artificial neural network classifier for classification of different stages. Three different kinds of eye disease conditions were tested in 350 subjects. We demonstrated a sensitivity of more than 95% for these classifiers with a specificity of 100%, and results are very promising. Our systems are ready to run clinically on large amounts of datasets.