ABSTRACT
Glaucoma is a chronic neurodegenerative disease that can result in irreversible vision loss if not treated in its early stages. The cup-to-disc ratio is a key criterion for glaucoma screening and diagnosis, and it is determined by dividing the area of the optic cup (OC) by that of the optic disc (OD) in fundus images. Consequently, the automatic and accurate segmentation of the OC and OD is a pivotal step in glaucoma detection. In recent years, numerous methods have resulted in great success on this task. However, most existing methods either have unsatisfactory segmentation accuracy or high time costs. In this paper, we propose a lightweight deep-learning architecture for the simultaneous segmentation of the OC and OD, where we have adopted fuzzy learning and a multi-layer perceptron to simplify the learning complexity and improve segmentation accuracy. Experimental results demonstrate the superiority of our proposed method as compared to most state-of-the-art approaches in terms of both training time and segmentation accuracy.
Subject(s)
Algorithms , Deep Learning , Fuzzy Logic , Glaucoma , Optic Disk , Humans , Optic Disk/diagnostic imaging , Glaucoma/diagnostic imaging , Neural Networks, Computer , Image Processing, Computer-Assisted/methods , Reproducibility of Results , Image Interpretation, Computer-Assisted/methods , Fundus OculiABSTRACT
The optic cup (OC) and optic disc (OD) are two critical structures in retinal fundus images, and their relative positions and sizes are essential for effectively diagnosing eye diseases. With the success of deep learning in computer vision, deep learning-based segmentation models have been widely used for joint optic cup and disc segmentation. However, there are three prominent issues that impact the segmentation performance. First, significant differences among datasets collecting from various institutions, protocols, and devices lead to performance degradation of models. Second, we find that images with only RGB information struggle to counteract the interference caused by brightness variations, affecting color representation capability. Finally, existing methods typically ignored the edge perception, facing the challenges in obtaining clear and smooth edge segmentation results. To address these drawbacks, we propose a novel framework based on Style Alignment and Multi-Color Fusion (SAMCF) for joint OC and OD segmentation. Initially, we introduce a domain generalization method to generate uniformly styled images without damaged image content for mitigating domain shift issues. Next, based on multiple color spaces, we propose a feature extraction and fusion network aiming to handle brightness variation interference and improve color representation capability. Lastly, an edge aware loss is designed to generate fine edge segmentation results. Our experiments conducted on three public datasets, DGS, RIM, and REFUGE, demonstrate that our proposed SAMCF achieves superior performance to existing state-of-the-art methods. Moreover, SAMCF exhibits remarkable generalization ability across multiple retinal fundus image datasets, showcasing its outstanding generality.
Subject(s)
Deep Learning , Optic Disk , Humans , Optic Disk/diagnostic imaging , Color , Algorithms , Image Interpretation, Computer-Assisted/methods , Image Processing, Computer-Assisted/methodsABSTRACT
Accurate image segmentation plays a crucial role in computer vision and medical image analysis. In this study, we developed a novel uncertainty guided deep learning strategy (UGLS) to enhance the performance of an existing neural network (i.e., U-Net) in segmenting multiple objects of interest from images with varying modalities. In the developed UGLS, a boundary uncertainty map was introduced for each object based on its coarse segmentation (obtained by the U-Net) and then combined with input images for the fine segmentation of the objects. We validated the developed method by segmenting optic cup (OC) regions from color fundus images and left and right lung regions from Xray images. Experiments on public fundus and Xray image datasets showed that the developed method achieved a average Dice Score (DS) of 0.8791 and a sensitivity (SEN) of 0.8858 for the OC segmentation, and 0.9605, 0.9607, 0.9621, and 0.9668 for the left and right lung segmentation, respectively. Our method significantly improved the segmentation performance of the U-Net, making it comparable or superior to five sophisticated networks (i.e., AU-Net, BiO-Net, AS-Net, Swin-Unet, and TransUNet).
ABSTRACT
Semi-supervised segmentation plays an important role in computer vision and medical image analysis and can alleviate the burden of acquiring abundant expert-annotated images. In this paper, we developed a residual-driven semi-supervised segmentation method (termed RDMT) based on the classical mean teacher (MT) framework by introducing a novel model-level residual perturbation and an exponential Dice (eDice) loss. The introduced perturbation was integrated into the exponential moving average (EMA) scheme to enhance the performance of the MT, while the eDice loss was used to improve the detection sensitivity of a given network to object boundaries. We validated the developed method by applying it to segment 3D Left Atrium (LA) and 2D optic cup (OC) from the public LASC and REFUGE datasets based on the V-Net and U-Net, respectively. Extensive experiments demonstrated that the developed method achieved the average Dice score of 0.8776 and 0.7751, when trained on 10% and 20% labeled images, respectively for the LA and OC regions depicted on the LASC and REFUGE datasets. It significantly outperformed the MT and can compete with several existing semi-supervised segmentation methods (i.e., HCMT, UAMT, DTC and SASS).
ABSTRACT
Some astronauts on International Space Station missions experience neuroophthalmological pathologies as part of spaceflight associated neuro-ocular syndrome (SANS). Strict head-down tilt bed rest (HDTBR) is a spaceflight analog that replicates SANS findings and those who had 3-4 risk alleles (G and C alleles from the methionine synthase reductase [MTRR] A66G and serine hydroxymethyltransferase [SHMT1] C1420T, respectively) as compared to 1-2 risk alleles, had a greater increase in total retinal thickness (TRT). The objective of this study was to identify factors that contribute to the individual variability of the development of SANS in a 60 d HDTBR at the German Aerospace Center's:envihab facility, Cologne Germany. 22 of 24 subjects who participated in the HDTBR study provided blood samples for genetic analysis. Total retinal thickness and optic cup volume were measured before and after bed rest. Subjects with 3-4 versus 0-2 risk alleles had greater ΔTRT during and after bed rest, and the model improved with the addition of baseline optic cup volume. This bed rest study confirms that variants of MTRR and SHMT1 are associated with ocular pathologies. Subjects with more risk alleles had the greatest HDTBR-induced ΔTRT, reaffirming that genetics predispose some individuals to developing SANS. Preflight optic cup volume and genetics better predict ΔTRT than either one alone. Whether nutritional supplements can override the genetic influences on biochemistry, physiology, and pathophysiology remains to be tested. These findings have significant implications for both aerospace and terrestrial medicine.
ABSTRACT
Glaucoma is one of blind causing diseases. The cup-to-disc ratio is the main basis for glaucoma screening. Therefore, it is of great significance to precisely segment the optic cup and disc. In this article, an optic cup and disc segmentation model based on the linear attention and dual attention is proposed. Firstly, the region of interest is located and cropped according to the characteristics of the optic disc. Secondly, linear attention residual network-34 (ResNet-34) is introduced as a feature extraction network. Finally, channel and spatial dual attention weights are generated by the linear attention output features, which are used to calibrate feature map in the decoder to obtain the optic cup and disc segmentation image. Experimental results show that the intersection over union of the optic disc and cup in Retinal Image Dataset for Optic Nerve Head Segmentation (DRISHTI-GS) dataset are 0.962 3 and 0.856 4, respectively, and the intersection over union of the optic disc and cup in retinal image database for optic nerve evaluation (RIM-ONE-V3) are 0.956 3 and 0.784 4, respectively. The proposed model is better than the comparison algorithm and has certain medical value in the early screening of glaucoma. In addition, this article uses knowledge distillation technology to generate two smaller models, which is beneficial to apply the models to embedded device.
Subject(s)
Humans , Optic Disk/diagnostic imaging , Glaucoma/diagnosis , Algorithms , Diagnostic Techniques, Ophthalmological , Databases, FactualABSTRACT
Glaucoma is the leading cause of irreversible blindness, but its early symptoms are not obvious and are easily overlooked, so early screening for glaucoma is particularly important. The cup to disc ratio is an important indicator for clinical glaucoma screening, and accurate segmentation of the optic cup and disc is the key to calculating the cup to disc ratio. In this paper, a full convolutional neural network with residual multi-scale convolution module was proposed for the optic cup and disc segmentation. First, the fundus image was contrast enhanced and polar transformation was introduced. Subsequently, W-Net was used as the backbone network, which replaced the standard convolution unit with the residual multi-scale full convolution module, the input port was added to the image pyramid to construct the multi-scale input, and the side output layer was used as the early classifier to generate the local prediction output. Finally, a new multi-tag loss function was proposed to guide network segmentation. The mean intersection over union of the optic cup and disc segmentation in the REFUGE dataset was 0.904 0 and 0.955 3 respectively, and the overlapping error was 0.178 0 and 0.066 5 respectively. The results show that this method not only realizes the joint segmentation of cup and disc, but also improves the segmentation accuracy effectively, which could be helpful for the promotion of large-scale early glaucoma screening.
Subject(s)
Humans , Diagnostic Techniques, Ophthalmological , Fundus Oculi , Glaucoma/diagnostic imaging , Neural Networks, Computer , Optic Disk/diagnostic imagingABSTRACT
PURPOSE: To compare parameters of confocal scanning laser ophthalmoscopy (Heidelberg Retinal Tomograph [HRT] II) in high myopia with age- and sex-matched emmetropes. METHODS: A hospital-based cross-sectional study was conducted among Nepalese subjects aged 18 to 35 years at BP Koirala Lions Centre for Ophthalmic Studies, Institute of Medicine, Maharajgunj, Kathmandu from November 2015 to October 2016. Fifty consecutive subjects with high myopia (spherical equivalent ranging from −6.00 to −12.00 diopters) and age- and sex-matched emmetropic subjects were enrolled for comparison. Correlations between disc area and other HRT parameters, asymmetry between the right and left eyes, and comparisons between male and female subjects in both high myopic and emmetropic groups were evaluated. RESULTS: Disc area was not significantly (p = 0.11) larger in high myopic eyes than in emmetropic eyes. HRT parameters in highly myopic eyes involved smaller cup parameters and greater rim parameters compared with emmetropic eyes. Disc area was found to be significantly positively correlated with inter disc parameters and significantly negatively correlated with rim to disc area ratio in the high myopia group. Disc area and other intra-disc parameters showed significant correlations between right and left eyes in both high myopia and emmetropia, and no significant differences between males and females from a Nepalese population. CONCLUSIONS: Characteristics of HRT parameters in high myopic eyes involved smaller cup parameters and greater rim parameters compared with emmetropic eyes in a Nepalese population. The effect of disc area on HRT parameters differed significantly only in height variation contour by emmetropic eyes.
Subject(s)
Female , Humans , Male , Cross-Sectional Studies , Emmetropia , Lions , Myopia , Ophthalmoscopy , Optic Disk , Optic Nerve , RetinaldehydeABSTRACT
PURPOSE: To evaluate the diagnostic ability of the modified ISNT rule (disc rim thickness of the smaller of inferior and superior > the larger of nasal and temporal) for normal and glaucomatous eyes compared to the classic ISNT rule (disc rim thickness of inferior > superior > nasal > temporal). METHODS: Color stereo optic disc photographs of 113 normal subjects and 108 open angle glaucoma patients with early and moderate stage were morphometrically evaluated. The classic ISNT rule and the modified ISNT rule were assessed by masked evaluation of disc photographs at the 3, 6, 9 and 12 o'clock positions. RESULTS: Among normal subjects, 58 of 113 eyes (51.3%) were normal and in open angle glaucoma patients, 104 of 108 eyes (96.3%) were abnormal with the classic ISNT rule. Among normal subjects, 98 of 113 eyes (94.2%) were normal and in open angle glaucoma patients, 102 of 108 eyes (94.4%) were abnormal with the modified ISNT rule. The modified ISNT rule was more accurate than the classic ISNT rule in terms of Cohen's Kappa analysis used for discriminating between normal and glaucomatous eyes. CONCLUSIONS: The modified ISNT rule is useful for differentiating between normal and glaucomatous optic nerves and easily applied in clinical practice.
Subject(s)
Humans , Glaucoma, Open-Angle , Masks , Optic NerveABSTRACT
Objective To explore the distribution and features of the optic cup stem cells in embryonic rat at tailbud stage. Methods The distribution of optic cup stem cells in optic cup tissue in 12.5-embryonic-day-old rats was observed by immunohistochemistry. The separated cells from optic cup were cultured with serum-free media, and immunofluorescence technique was used to detect the ability of hyperplasia of stem cells and expression of CHX10 antigen and specific antigens of mature retinal cells before and after differentiation. Results The optic cup stem cells in embryonic rat at tailbud stage were mainly located at inner, outer, and marginal layer of optic cup. No expression of specifically marked protein of mature retinal cells was detected. The cells separated from optic cup had the ability of single-cell clone, positive expression of CHX10 and expression of several specific antigens of mature retinal cells after the inducement, including Thy1.1, glial fibrillary acidic protein (GFAP), protein kinase C (PKC) ?, and rhodopsin. Conclusion Optic cup of 12.5-embryonic-day-old rats composes of undifferentiated cells, and the stem cells are mainly located in optic cup inner and marginal. High ability of hyperplasia of the optic cup stem cells cultured in vitro is found. The cells, which are retinal stem cells, can express several specifically marked proteins of mature retinal cells after inducement and differentiation.
ABSTRACT
Objective To study the location of embryonic optic cup stem cells during tailbud stage. Methods The embryonic optic cup at embryonic day 11~15 (E11~15) in rats was sectioned horizontally at 15 ?m thick. The distributive characteristics of embryonic optic cup progenitor cells were revealed by immunohistochemistry. Results ①The distribution of optic cup progenitors was mainly aggregated on the optic cup at E12.5. CHX10-positive cells were organized as stratified epithelium arrangement on optic cup inner layer. Clusters of CHX10-positive cells were observed at the edge of optic cup; ② Pigment appeared in the outer layer of optic cup at E13.5, and differentiation into ganglion cells was initiated. Conclusion The distribution of optic cup stem cells is mainly aggregated on the optic cup at E12.5 in which the differentiation into ganglion cells is not initiated.
ABSTRACT
Three types of microscopic operations were used in this study: 1. excision of the optic vesicle alone; 2. after excising the optic vesicle, a piece of forebrain tissue taken from another donor was inserted in between presumptive lens ectoderm and forebrain wall; 3. cut down the optic vesicle and portion of forebrain tissue and replaced them back in situ by turning over 180?. The operations were carried out on Rana nigromaculata at 15 and 16 stages, in order to bring the presumptive lens ectoderm to come into contact directly with the forebrain tissue for the purpose to analyse the possibility of eye cup formation from the brain wall. The results indicate that the forebrain wall which came into contact with the presumptive lens ectoderm could be induced and differentiated into a secondary eye cup or retina. A total of 81 cases of secondary eye cups among 282 operations (28.8%) were observed. The mechanism for the induction and its significance were discussed.