2024-03-29T01:40:19Z
https://pubs.lib.uiowa.edu/omia/api/oai/
oai:omia:id:27639
2016-10-21T01:00:00Z
A Depth Based Approach to Glaucoma Detection Using Retinal Fundus Images
A Depth Based Approach to Glaucoma Detection Using Retinal Fundus Images
Ramaswamy, Akshaya
Ram, Keerthi
Sivaprakasam, Mohanasankar
Qualitative evaluation of stereo retinal fundus images by experts is a widely accepted method for optic nerve head evaluation (ONH) in glaucoma. The quantitative evaluation using stereo involves depth estimation of the ONH and thresholding of depth to extract optic cup. In this paper, we attempt the reverse, by estimating the disc depth using supervised and unsupervised techniques on a single optic disc image. Our depth estimation approach is evaluated on the INSPIRE-stereo dataset by using single images from the stereo pairs, and is compared with the OCT based depth ground truths. We extract spatial and intensity features from the depth maps, and perform classification of images into glaucomatous and normal. Our approach is evaluated on a dataset of 100 images and achieves an AUC of 0.888 with a sensitivity of 83% at specificity 83%. Experiments indicate that our approach can reliably estimate depth, and provide valuable information for glaucoma detection and for monitoring its progression.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1041
https://doi.org/10.17077/omia.1041
https://pubs.lib.uiowa.edu/omia/article/id/27639/
http://rightsstatements.org/vocab/InC/1.0/
9-16
oai:omia:id:27635
2016-10-21T01:00:00Z
A Novel Machine Learning Model Based on Exudate Localization to Detect Diabetic Macular Edema
A Novel Machine Learning Model Based on Exudate Localization to Detect Diabetic Macular Edema
Perdomo, Oscar
Otalora, Sebastian
Rodríguez, Francisco
Arevalo, John
González, Fabio A.
Diabetic macular edema is one of the leading causes of legal blindness worldwide. Early, and accessible, detection of ophthalmological diseases is especially important in developing countries, where there are major limitations to access to specialized medical diagnosis and treatment. Deep learning models, such as deep convolutional neural networks have shown great success in different computer vision tasks. In medical images they have been also applied with great success. The present paper presents a novel strategy based on convolutional neural networks to combine exudates localization and eye fundus images for automatic classification of diabetic macular edema as a support for diabetic retinopathy diagnosis.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1057
https://doi.org/10.17077/omia.1057
https://pubs.lib.uiowa.edu/omia/article/id/27635/
http://rightsstatements.org/vocab/InC/1.0/
137-144
oai:omia:id:27644
2016-10-21T01:00:00Z
Anterior Chamber Angle Assessment System
Anterior Chamber Angle Assessment System
Fu, Huazhu
Xu, Yanwu
Kee Wong, Damon Wing
Liu, Jiang
Baskaran, Mani
Perera, Shamira A.
Aung, Tin
In this paper, we propose an automatic anterior chamber angle assessment system for Anterior Segment Optical Coherence Tomography (AS-OCT). In our system, the automatic segmentation method is used to segment the clinical structures, which are then used to recover standard clinical ACA measurements. Our measurements can not only support clinical assessments, but also be utilized as features for detecting anterior angle closure in automatic glaucoma diagnosis.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1059
https://doi.org/10.17077/omia.1059
https://pubs.lib.uiowa.edu/omia/article/id/27644/
http://rightsstatements.org/vocab/InC/1.0/
152-153
oai:omia:id:27626
2016-10-21T01:00:00Z
Artefacts Removal from Optical Coherence Tomography Angiography
Artefacts Removal from Optical Coherence Tomography Angiography
Ong, Ee Ping
Cheng, Jun
Quan, Ying
Xu, Guozhen
Wong, Damon W.K.
This paper presents a new approach for artefacts removal from optical coherence tomography angiography (OCTA). The artefacts mainly arise as a result of distortion due to eye movements during OCT scanning process. These distortions manifest themselves as visible motion artefacts when doctors review the enface image of OCTA data. To remove these artefacts, firstly we perform motion registration for the captured OCT volume data and subsequently perform motion correction to obtain the registered OCT data. Next, we compute the OCTA from the registered OCT data using an enhanced correlation mapping technique. Thereafter, we compute the enface image from the OCTA data. In the next step, we attempt to locate regions where there is misalignment in the OCT frames of the various B-scans. Finally, we attempt to restore the regions where correct data is postulated to be absent. Our experimental results demonstrate the effectiveness of our proposed approach.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1049
https://doi.org/10.17077/omia.1049
https://pubs.lib.uiowa.edu/omia/article/id/27626/
http://rightsstatements.org/vocab/InC/1.0/
73-80
oai:omia:id:27627
2016-10-21T01:00:00Z
Automated Morphometric Analysis of in-vivo Human Corneal Endothelium
Automated Morphometric Analysis of in-vivo Human Corneal Endothelium
Scarpa, Fabio
Gassa, Chiara Dalla
Ruggeri, Alfredo
In-vivo specular and confocal microscopy provide information on the corneal endothelium health state. The reliable estimation of the clinical parameters requires the accurate detection of cell contours. We propose a method for the automatic segmentation of cell contour. The centers of the cells are detected by convolving the original image with Laplacian of Gaussian kernels, whose scales are set according to the cell size preliminary estimated through a frequency analysis. A structure made by connected vertices is derived from the centers, and it is fine-tuned by combining information about the typical regularity of endothelial cells shape with the pixels intensity of the actual image. Ground truth values for the clinical parameters were obtained from manually drawn cell contours. An accurate automatic estimation is achieved on 30 images: for each clinical parameter, the mean difference between its manual estimation and the automated one is always less than 7%.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1051
https://doi.org/10.17077/omia.1051
https://pubs.lib.uiowa.edu/omia/article/id/27627/
http://rightsstatements.org/vocab/InC/1.0/
89-96
oai:omia:id:27637
2016-10-21T01:00:00Z
Automated Tessellated Fundus Detection in Color Fundus Images
Automated Tessellated Fundus Detection in Color Fundus Images
Xu, Mengdi
Cheng, Jun
Kee Wong, Damon Wing
Cheng, Ching-Yu
Saw, Seang Mei
Wong, Tien Yin
In this work, we propose an automated tessellated fundus detection method by utilizing texture features and color features. Color moments, Local Binary Patterns (LBP), and Histograms of Oriented Gradients (HOG) are extracted to represent the color fundus image. After feature extraction, a SVM classifier is trained to detect the tessellated fundus. Both linear and RBF kernels are applied and compared in this work. A dataset with 836 fundus images is built to evaluate the proposed method. For linear SVM, the mean accuracy of 98% is achieved, with sensitivity of 0.99 and specificity of 0.98. For RBF kernel, the mean accuracy is 97%, with sensitivity of 0.99 and specificity of 0.95. The detection results indicate that color features and texture features are able to describe the tessellated fundus.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1043
https://doi.org/10.17077/omia.1043
https://pubs.lib.uiowa.edu/omia/article/id/27637/
http://rightsstatements.org/vocab/InC/1.0/
25-32
oai:omia:id:27640
2016-10-21T01:00:00Z
Automatic Optic Disc Abnormality Detection in Fundus Images: A Deep Learning Approach
Automatic Optic Disc Abnormality Detection in Fundus Images: A Deep Learning Approach
Alghamdi, Hanan S.
Tang, Hongying Lilian
Waheeb, Saad A.
Peto, Tunde
Optic disc (OD) is a key structure in retinal images. It serves as an indicator to detect various diseases such as glaucoma and changes related to new vessel formation on the OD in diabetic retinopathy (DR) or retinal vein occlusion. OD is also essential to locate structures such as the macula and the main vascular arcade. Most existing methods for OD localization are rule-based, either exploiting the OD appearance properties or the spatial relationship between the OD and the main vascular arcade. The detection of OD abnormalities has been performed through the detection of lesions such as hemorrhaeges or through measuring cup to disc ratio. Thus these methods result in complex and inflexible image analysis algorithms limiting their applicability to large image sets obtained either in epidemiological studies or in screening for retinal or optic nerve diseases. In this paper, we propose an end-to-end supervised model for OD abnormality detection. The most informative features of the OD are learned directly from retinal images and are adapted to the dataset at hand. Our experimental results validated the effectiveness of this current approach and showed its potential application.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1042
https://doi.org/10.17077/omia.1042
https://pubs.lib.uiowa.edu/omia/article/id/27640/
http://rightsstatements.org/vocab/InC/1.0/
17-24
oai:omia:id:27642
2016-10-21T01:00:00Z
Bridging Disconnected Curvilinear Structures via Numerical Evolutions of Completion Process in Ophthalmologic Images
Bridging Disconnected Curvilinear Structures via Numerical Evolutions of Completion Process in Ophthalmologic Images
Zhang, Jiong
Bekkers, Erik
Abbasi-Sureshjani, Samaneh
Dashtbozorg, Behdad
Haar Romeny, Bart ter
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1061
https://doi.org/10.17077/omia.1061
https://pubs.lib.uiowa.edu/omia/article/id/27642/
http://rightsstatements.org/vocab/InC/1.0/
156-157
oai:omia:id:27641
2016-10-21T01:00:00Z
Diabetic Macular Edema Grading Based on Deep Neural Networks
Diabetic Macular Edema Grading Based on Deep Neural Networks
Al-Bander, Baidaa
Al-Nuaimy, Waleed
Al-Taee, Majid A.
Williams, Bryan M.
Zheng, Yalin
Diabetic Macular Edema (DME) is a major cause of vision loss in diabetes. Its early detection and treatment is therefore a vital task in management of diabetic retinopathy. In this paper, we propose a new featurelearning approach for grading the severity of DME using color retinal fundus images. An automated DME diagnosis system based on the proposed featurelearning approach is developed to help early diagnosis of the disease and thus averts (or delays) its progression. It utilizes the convolutional neural networks (CNNs) to identify and extract features of DME automatically without any kind of user intervention. The developed prototype was trained and assessed by using an existing MESSIDOR dataset of 1200 images. The obtained preliminary results showed accuracy of (88.8 %), sensitivity (74.7%) and specificity (96.5 %). These results compare favorably to state-of-the-art findings with the added benefit of an automatic feature-learning approach rather than a time-consuming handcrafted approach.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1055
https://doi.org/10.17077/omia.1055
https://pubs.lib.uiowa.edu/omia/article/id/27641/
http://rightsstatements.org/vocab/InC/1.0/
121-128
oai:omia:id:27643
2016-10-21T01:00:00Z
Evaluation of the Areas Involved in Visual Cortex in Parkinson's Disease Using Diffusion Tensor Imaging
Evaluation of the Areas Involved in Visual Cortex in Parkinson's Disease Using Diffusion Tensor Imaging
Jooyandeh, Somayeh Mohammadi
Kamalian, Aida
Shiranvand, Sepideh
Dolatshahi, Mahsa
Shadmehr, Mohammad Hadi
Baghai, Thomas C.
Rahmani, Farzaneh
Shojaie, Ahmad
Aarabi, Mohammad Hadi
Parkinson's disease (PD) is a progressive neurodegenerative disorder assumed to involve different areas of CNS and PNS. In PD patients and in primates with experimental Parkinsonism indicating that retinal dopamine deficiency is an important factor in the pathogenesis of PD visual dysfunction. Visual signs and symptoms of PD may include defects in eye movement, pupillary function, and in more complex visual tasks. In this study, we evaluated the areas involved in visual cortex in PD by diffusion tensor imaging to assess the structural change in PD.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1058
https://doi.org/10.17077/omia.1058
https://pubs.lib.uiowa.edu/omia/article/id/27643/
http://rightsstatements.org/vocab/InC/1.0/
145-151
oai:omia:id:27636
2016-10-21T01:00:00Z
Geometric Connectivity Analysis Based on Edge Co-Occurrences in Retinal Images
Geometric Connectivity Analysis Based on Edge Co-Occurrences in Retinal Images
Abbasi-Sureshjani, Samaneh
Zhang, Jiong
Sanguinetti, Gonzalo
Duits, Remco
Haar Romeny, Bart ter
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1060
https://doi.org/10.17077/omia.1060
https://pubs.lib.uiowa.edu/omia/article/id/27636/
http://rightsstatements.org/vocab/InC/1.0/
154-155
oai:omia:id:27633
2016-10-21T01:00:00Z
Image Quality Classification for DR Screening Using Convolutional Neural Networks
Image Quality Classification for DR Screening Using Convolutional Neural Networks
Tennakoon, Ruwan
Mahapatra, Dwarikanath
Roy, Pallab
Sedai, Suman
Garnavi, Rahil
The quality of input images significantly affects the outcome of automated diabetic retinopathy screening systems. Current methods to identify image quality rely on hand-crafted geometric and structural features, that does not generalize well. We propose a new method for retinal image quality classification (IQC) that uses computational algorithms imitating the working of the human visual systems. The proposed method leverages on learned supervised information using convolutional neural networks (CNN), thus avoiding hand-engineered features. Our analysis shows that the learned features capture both geometric and structural information relevant for image quality classification. Experimental results conducted on a relatively large dataset demonstrates that the overall method can achieve high accuracy. We also show that effective features for IQC can be learned by full training of shallow CNN as well as by using transfer learning.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1054
https://doi.org/10.17077/omia.1054
https://pubs.lib.uiowa.edu/omia/article/id/27633/
http://rightsstatements.org/vocab/InC/1.0/
113-120
oai:omia:id:27629
2016-10-21T01:00:00Z
Infrastructure for Retinal Image Analysis
Infrastructure for Retinal Image Analysis
Dashtbozorg, Behdad
Abbasi-Sureshjani, Samaneh
Zhang, Jiong
Huang, Fan
Bekkers, Erik
Haar Romeny, Bart ter
This paper introduces a retinal image analysis infrastructure for the automatic assessment of biomarkers related to early signs of diabetes, hypertension and other systemic diseases. The developed application provides several tools, namely normalization, vessel enhancement and segmentation, optic disc and fovea detection, junction detection, bifurcation/crossing discrimination, artery/vein classification and red lesion detection. The pipeline of these methods allows the assessment of important biomarkers characterizing dynamic properties of retinal vessels, such as tortuosity, width, fractal dimension and bifurcation geometry features.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1053
https://doi.org/10.17077/omia.1053
https://pubs.lib.uiowa.edu/omia/article/id/27629/
http://rightsstatements.org/vocab/InC/1.0/
105-112
oai:omia:id:27631
2016-10-21T01:00:00Z
Intensity-based Choroidal Registration Using Regularized Block Matching
Intensity-based Choroidal Registration Using Regularized Block Matching
Ronchetti, Tiziano
Maloca, Peter
Meier, Christoph
Orgül, Selim
Jud, Christoph
Hasler, Pascal
Považay, Boris
Cattin, Philippe C.
Detecting and monitoring changes in the human choroid play a crucial role in treating ocular diseases such as myopia. However, reliable segmentation of optical coherence tomography (OCT) images at the choroid-sclera interface (CSI) is notoriously difficult due to poor contrast, signal loss and OCT artefacts. In this paper we present blockwise registration of successive scans to improve stability also during complete loss of the CSI-signal. First, we formulated the problem as minimization of a regularized energy functional. Then, we tested our automated method for piecewise Intensity-based Choroidal rigid Registration using regularized block matching (ICR) on 20 OCT 3D-volume scan-rescan data set pairs. Finally, we used these data set pairs to determine the precision of our method, while the accuracy was determined by comparing our results with those using manually annotated scans.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1044
https://doi.org/10.17077/omia.1044
https://pubs.lib.uiowa.edu/omia/article/id/27631/
http://rightsstatements.org/vocab/InC/1.0/
33-40
oai:omia:id:27624
2016-10-21T01:00:00Z
Motion Correction in Optical Coherence Tomography for Multi-modality Retinal Image Registration
Motion Correction in Optical Coherence Tomography for Multi-modality Retinal Image Registration
Cheng, Jun
Lee, Jimmy Addison
Xu, Guozhen
Quan, Ying
Ong, Ee Ping
Kee Wong, Damon Wing
Optical coherence tomography (OCT) is a recently developed non-invasive imaging modality, which is often used in ophthalmology. Because of the sequential scanning in form of A-scans, OCT suffers from the inevitable eye movement. This often leads to mis-alignment especially among consecutive B-scans, which affects the analysis and processing of the data such as the registration of the OCT en face image to color fundus image. In this paper, we propose a novel method to correct the mis-alignment among consecutive B-scans to improve the accuracy in multi-modality retinal image registration. In the method, we propose to compute decorrelation from overlapping B-scans and to detect the eye movement. Then, the B-scans with eye movement will be re-aligned to its precedent scans while the rest of B-scans without eye movement are untouched. Our experiments results show that the proposed method improves the accuracy and success rate in the registration to color fundus images.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1048
https://doi.org/10.17077/omia.1048
https://pubs.lib.uiowa.edu/omia/article/id/27624/
http://rightsstatements.org/vocab/InC/1.0/
65-72
oai:omia:id:27628
2016-10-21T01:00:00Z
Optic Cup Segmentation Using Large Pixel Patch Based CNNs
Optic Cup Segmentation Using Large Pixel Patch Based CNNs
Guo, Yundi
Zou, Beiji
Chen, Zailiang
He, Qi
Liu, Qing
Zhao, Rongchang
Optic cup(OC) segmentation on color fundus image is essential for the calculation of cup-to-disk ratio and fundus morphological analysis, which are very important references in the diagnosis of glaucoma. In this paper we proposed an OC segmentation method using convolutional neural networks(CNNs) to learn from big size patch belong to each pixel. The segmentation result is achieved by classification of each pixel patch and postprocessing. With large pixel patch, the network could learn more global information around each pixel and make a better judgement during classification. We tested this method on public dataset Drishti-GS and achieved average F-Score of 93.73% and average overlapping error of 12.25%, which is better than state-of-the-art algorithms. This method could be used for fundus morphological analysis, and could also be employed to other medical image segmentation works which the boundary of the target area is fuzzy.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1056
https://doi.org/10.17077/omia.1056
https://pubs.lib.uiowa.edu/omia/article/id/27628/
http://rightsstatements.org/vocab/InC/1.0/
129-136
oai:omia:id:27625
2016-10-21T01:00:00Z
Predicting Drusen Regression from OCT in Patients with Age-Related Macular Degeneration
Predicting Drusen Regression from OCT in Patients with Age-Related Macular Degeneration
Bogunović, Hrvoje
Montuoro, Alessio
Waldstein, Sebastian M.
Baratsits, Magdalena
Schlanitz, Ferdinand
Schmidt-Erfurth, Ursula
Age-related macular degeneration (AMD) is a leading cause of blindness in developed countries. The presence of drusen is the hallmark of early/intermediate AMD, and their sudden regression is strongly associated with the onset of late AMD. In this work we propose a predictive model of drusen regression using optical coherence tomography (OCT) based features. First, a series of automated image analysis steps are applied to segment and characterize individual drusen and their development. Second, from a set of quantitative features, a random forest classifiser is employed to predict the occurrence of individual drusen regression within the following 12 months. The predictive model is trained and evaluated on a longitudinal OCT dataset of 44 eyes from 26 patients using leave-one-patient-out cross-validation. The model achieved an area under the ROC curve of 0.81, with a sensitivity of 0.74 and a specificity of 0.73. The presence of hyperreflective foci and mean drusen signal intensity were found to be the two most important features for the prediction. This preliminary study shows that predicting drusen regression is feasible and is a promising step toward identification of imaging biomarkers of incoming regression.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1045
https://doi.org/10.17077/omia.1045
https://pubs.lib.uiowa.edu/omia/article/id/27625/
http://rightsstatements.org/vocab/InC/1.0/
41-48
oai:omia:id:27634
2016-10-21T01:00:00Z
Restoration of Neonatal Retinal Images
Restoration of Neonatal Retinal Images
Shankaranarayana, Sharath M.
Ram, Keerthi
Vinekar, Anand
Mitra, Kaushik
Sivaprakasam, Mohanasankar
Retinopathy of prematurity (ROP) is an eye disorder primarily affecting premature neonates. Specialists use a number of neonatal retinal images acquired by a wide field of view camera for diagnosis and the subsequent follow up. However, the premature infants’ retinal images are generally of lower visibility compared to adult retinal images, affecting the quality of diagnosis. We study some image dehazing methods from general outdoor scenes and propose an image restoration scheme for neonatal retinal images, based on the physical model of light propagation in a medium. The results from our restoration algorithm is useful for analysis by human experts as well as computer aided diagnosis and specifically we show that our method enhances vessel segmentation significantly compared to traditional methods like adaptive histogram equalization.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1046
https://doi.org/10.17077/omia.1046
https://pubs.lib.uiowa.edu/omia/article/id/27634/
http://rightsstatements.org/vocab/InC/1.0/
49-56
oai:omia:id:27632
2016-10-21T01:00:00Z
Retinal Image Quality Classification Using Neurobiological Models of the Human Visual System
Retinal Image Quality Classification Using Neurobiological Models of the Human Visual System
Mahapatra, Dwarikanath
Retinal image quality assessment (IQA) algorithms use different hand crafted features without considering the important role of the human visual system (HVS). We solve the IQA problem using the principles behind the working of the HVS. Unsupervised information from local saliency maps and supervised information from trained convolutional neural networks (CNNs) are combined to make a final decision on image quality. A novel algorithm is proposed that calculates saliency values for every image pixel at multiple scales to capture global and local image information. This extracts generalized image information in an unsupervised manner while CNNs provide a principled approach to feature learning without the need to define hand-crafted features. The individual classification decisions are fused by weighting them according to their confidence scores. Experimental results on real datasets demonstrate the superior performance of our proposed algorithm over competing methods.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1052
https://doi.org/10.17077/omia.1052
https://pubs.lib.uiowa.edu/omia/article/id/27632/
http://rightsstatements.org/vocab/InC/1.0/
97-104
oai:omia:id:27630
2016-10-21T01:00:00Z
Retinal Vessel Segmentation from Simple to Difficult
Retinal Vessel Segmentation from Simple to Difficult
Liu, Qing
Zou, Beiji
Chen, Jie
Chen, Zailiang
Zhu, Chengzhang
Yue, Kejuan
Zhao, Guoying
In this paper, we propose two vesselness maps and a simple to difficult learning framework for retinal vessel segmentation which is ground truth free. The first vesselness map is the multiscale centrelineboundary contrast map which is inspired by the appearance of vessels. The other is the difference of diffusion map which measures the difference of the diffused image and the original one. Meanwhile, two existing vesselness maps are generated. Totally, 4 vesselness maps are generated. In each vesselness map, pixels with large vesselness values are regarded as positive samples. Pixels around the positive samples with small vesselness values are regarded as negative samples. Then we learn a strong classifier for the retinal image based on other 3 vesselness maps to determine the pixels with mediocre values in single vesselness map. Finally, pixels with two classifier supports are labelled as vessel pixels. The experimental results on DRIVE and STARE show that our method outperforms the state-of-the-art unsupervised methods and achieves competitive performances to supervised methods.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1047
https://doi.org/10.17077/omia.1047
https://pubs.lib.uiowa.edu/omia/article/id/27630/
http://rightsstatements.org/vocab/InC/1.0/
57-64
oai:omia:id:27623
2016-10-21T01:00:00Z
Segmentation of Optic Disc and Optic Cup in Retinal Fundus Images Using Coupled Shape Regression
Segmentation of Optic Disc and Optic Cup in Retinal Fundus Images Using Coupled Shape Regression
Sedai, Suman
Roy, Pallab
Mahapatra, Dwarikanath
Garnavi, Rahil
Accurate segmentation of optic cup and disc in retinal fundus images is required to derive the cup-to-disc ratio (CDR) parameter which is the main indicator for Glaucoma assessment. In this paper, we propose a coupled regression method for accurate segmentation of optic cup and disc in retinal colour fundus image. The proposed coupled regression framework consists of a parameter regressor which directly predicts CDR from a given image, as well as an ensemble shape regressor which iteratively estimates the OD-OC boundary by taking into account the CDR estimated by the parameter regressor. The parameter regressor and the shape regressor are then coupled together within a feedback loop so that estimation of one reinforces the other. Both parameter regressor and the ensemble shape regressor are modeled using Boosted Regression Trees. The proposed optic cup and disc segmentation method is applied on an image set of 50 patients and demonstrates high segmentation accuracy. A comparative study shows that our proposed method outperforms state of the art methods for cup segmentation.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1040
https://doi.org/10.17077/omia.1040
https://pubs.lib.uiowa.edu/omia/article/id/27623/
http://rightsstatements.org/vocab/InC/1.0/
1-8
oai:omia:id:27622
2016-10-21T01:00:00Z
Stereo Eye Tracking with a Single Camera for Ocular Tumor Therapy
Stereo Eye Tracking with a Single Camera for Ocular Tumor Therapy
Wyder, Stephan
Cattin, Philippe C.
We present a compact and accurate stereo eye tracking system using only one physical camera. The proposed eye tracking system is intended as a navigation system for ocular tumor therapy. There, the available physical space to mount an eye tracker is limited. Furthermore, high system accuracy is demanded. However, high eye tracker accuracy and system compactness often disagree. Current established eye trackers can live with that compromise, desktop devices focus more on accuracy whereas mobile devices focus on compactness. We combine a stereo eye tracking algorithm with a clever arrangement of two planar mirrors and a single camera to get high accuracy, precision and a compact design altogether. We developed an eye tracking prototype and tested the system with ten healthy volunteers. We show that the proposed eye tracker is more accurate and robust, while at the same time equally compact as a comparable eye tracking system containing one instead of two mirrors.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1050
https://doi.org/10.17077/omia.1050
https://pubs.lib.uiowa.edu/omia/article/id/27622/
http://rightsstatements.org/vocab/InC/1.0/
81-88
oai:omia:id:27638
2016-10-21T01:00:00Z
Vessel Extraction for AS-OCT Angiography
Vessel Extraction for AS-OCT Angiography
Fu, Huazhu
Xu, Yanwu
Kee Wong, Damon Wing
Ang, Marcus
Das, Suchandrima
Liu, Jiang
In this work, we propose a filter-based vessel segmentation method for Anterior Segment Optical Coherence Tomography Angiography image. In our method, the bandpass filter is utilized to suppress the horizontal noise lines caused by eye movement, while the curvedsupport Gaussian filter is utilized to enhance the vessel and generate the probability map.
2016-10-21T01:00:00Z
info:eu-repo/semantics/conferenceObject
3
2016
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1062
https://doi.org/10.17077/omia.1062
https://pubs.lib.uiowa.edu/omia/article/id/27638/
http://rightsstatements.org/vocab/InC/1.0/
158-159
oai:omia:id:27663
2015-10-09T01:00:00Z
Adaptive Super-Candidate Based Approach for Detection and Classification of Drusen on Retinal Fundus Images
Adaptive Super-Candidate Based Approach for Detection and Classification of Drusen on Retinal Fundus Images
Sundaresan, Vaanathi
Ram, Keerthi
Selvaraj, Kulasekaran
Joshi, Niranjan
Sivaprakasam, Mohanasankar
Identification and characterization of drusen is essential for the severity assessment of age-related macular degeneration (AMD). Presented here is a novel super-candidate based approach, combined with robust preprocessing and adaptive thresholding for detection of drusen, resulting in accurate segmentation with the mean lesion-level overlap of 0.75, even in cases with non-uniform illumination, poor contrast and con- founding anatomical structures. We also present a feature based lesion- level discrimination analysis between hard and soft drusen. Our method gives sensitivity of 80% for high specificity above 90% and high sensitivity of 95% for specificity of 70% on representative pathological databases (STARE and ARIA) for both detection and discrimination.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1030
https://doi.org/10.17077/omia.1030
https://pubs.lib.uiowa.edu/omia/article/id/27663/
http://rightsstatements.org/vocab/InC/1.0/
81-88
oai:omia:id:27669
2015-10-09T01:00:00Z
A New Method of Blind Deconvolution for Colour Fundus Retinal Images
A New Method of Blind Deconvolution for Colour Fundus Retinal Images
Williams, Bryan M.
Chen, Ke
Harding, Simon P.
Zheng, Yalin
Fundus retinal imaging is widely used in the diagnosis and management of eye disease. Blur commonly occurs in the acquisition and when it is severe the resulting loss of resolution hampers accurate clinical assessment. In this paper, we present a new technique to address this challenging problem. We make use of implicitly constrained image deblurring, which is known to provide improved results over unconstrained and explicitly constrained methods, and build this into a multi-channel variational framework for parametric deblurring. We propose a new method for automatically selecting the regularisation parameter in the absence of the true (sharp) image using vessel segmentation. We then modify the model to include a regularisation coefficient function which is dependent on an available image mask in order to avoid potential inaccuracies caused by the addition of artificial masks. We present experimental results to demonstrate the effectiveness of our new method.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1036
https://doi.org/10.17077/omia.1036
https://pubs.lib.uiowa.edu/omia/article/id/27669/
http://rightsstatements.org/vocab/InC/1.0/
129-136
oai:omia:id:27681
2015-10-09T01:00:00Z
A Polar Map Based Approach Using Retinal Fundus Images for Glaucoma Detection
A Polar Map Based Approach Using Retinal Fundus Images for Glaucoma Detection
Ramaswamy, Akshaya
Ram, Keerthi
Joshi, Niranjan
Sivaprakasam, Mohanasankar
Cup-to-disc ratio is commonly used as an important parameter for glaucoma screening, involving segmentation of the optic cup on fundus images. We propose a novel polar map representation of the optic disc, using a combination of supervised and unsupervised cup segmentation techniques, for detection of glaucoma. Instead of performing hard thresholding on the segmentation output to extract the cup, we consider the cup confidence scores inside the disc to construct a polar map, and extract sector-wise features for learning a glaucoma risk probability (GRP) for the image. We compare the performance of GRP vis-à-vis the cup-to-disc ratio (CDR). On an evaluation dataset of 100 images from the publicly available RIM-ONE database, our method achieves 82% sensitivity at 84% specificity, and 96% sensitivity at 60% specificity (AUC of 0.8964). Experiments indicate that the polar map based method can provide a more discriminatory glaucoma risk probability score compared to CDR.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1038
https://doi.org/10.17077/omia.1038
https://pubs.lib.uiowa.edu/omia/article/id/27681/
http://rightsstatements.org/vocab/InC/1.0/
145-152
oai:omia:id:27672
2015-10-09T01:00:00Z
Automated Bruch’s Membrane Opening Segmentation in Cases of Optic Disc Swelling in Combined 2D and 3D SD-OCT Images Using Shape-Prior and Texture Information
Automated Bruch’s Membrane Opening Segmentation in Cases of Optic Disc Swelling in Combined 2D and 3D SD-OCT Images Using Shape-Prior and Texture Information
Wang, Jui-Kai
Kardon, Randy H.
Garvin, Mona K.
When the optic disc is swollen, the visibility of the Bruch’s membrane opening (BMO) is often drastically reduced in spectral-domain optical coherence tomography (SD-OCT) volumes. Recent work pro- posed a semi-automated method to segment the BMO using combined information from 2D high-definition raster and 3D volumetric SD-OCT scans; however, manual placement of six landmark points was required. In this work, we propose a fully automated approach to segment the BMO from 2D high-definition and 3D volumetric SD-OCT scans. Using the topographic shape of the internal limiting membrane and textural information near Bruch’s membrane, two BMO points are first estimated in the high-definition central B-scan and then registered into the corresponding volumetric scan. Utilizing the information from both the high- definition BMO estimates and the standard-definition SD-OCT volume, the cost image was created. A graph-based algorithm with soft shape-based constraints is further applied to segment the BMO contour on the SD-OCT en-face image domain. Using a set of 23 volumes with reasonably centered raster scans and swelling larger than 14.42 mm3, the fully automated approach was significantly more accurate than a traditional approach utilizing information only from the SD-OCT volume (RMS error of 7.18 vs. 21.37 in pixels; p < 0.05) and had only a slightly higher (and not significantly different) error than the previously proposed semi-automated approach (RMS error of 7.18 vs. 5.30 in pixels; p = 0.08).
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1024
https://doi.org/10.17077/omia.1024
https://pubs.lib.uiowa.edu/omia/article/id/27672/
http://rightsstatements.org/vocab/InC/1.0/
33-40
oai:omia:id:27673
2015-10-09T01:00:00Z
Automatic Grading of Diabetic Retinopathy on a Public Database
Automatic Grading of Diabetic Retinopathy on a Public Database
Seoud, Lama
Chelbi, Jihed
Cheriet, Farida
With the growing diabetes epidemic, retina specialists have to examine a tremendous amount of fundus images for the detection and grading of diabetic retinopathy. In this study, we propose a first automatic grading system for diabetic retinopathy. First, a red lesion detection is performed to generate a lesion probability map. The latter is then represented by 35 features combining location, size and probability information, which are finally used for classification. A leave-one-out cross-validation using a random forest is conducted on a public database of 1200 images, to classify the images into 4 grades. The proposed system achieved a classification accuracy of 74.1% and a weighted kappa value of 0.731 indicating a significant agreement with the reference. These preliminary results prove that automatic DR grading is feasible, with a performance comparable to that of human experts.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1032
https://doi.org/10.17077/omia.1032
https://pubs.lib.uiowa.edu/omia/article/id/27673/
http://rightsstatements.org/vocab/InC/1.0/
97-104
oai:omia:id:27682
2015-10-09T01:00:00Z
Boosting Convolutional Filters with Entropy Sampling for Optic Cup and Disc Image Segmentation from Fundus Images
Boosting Convolutional Filters with Entropy Sampling for Optic Cup and Disc Image Segmentation from Fundus Images
Zilly, Julian G.
Buhmann, Joachim M.
Mahapatra, Dwarikanath
We propose a novel convolutional neural network (CNN) based method for optic cup and disc segmentation. To reduce computational complexity, an entropy based sampling technique is introduced that gives superior results over uniform sampling. Filters are learned over several layers with the output of previous layers serving as the input to the next layer. A softmax logistic regression classifier is subsequently trained on the output of all learned filters. In several error metrics, the proposed algorithm outperforms existing methods on the public DRISHTI-GS data set.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1039
https://doi.org/10.17077/omia.1039
https://pubs.lib.uiowa.edu/omia/article/id/27682/
http://rightsstatements.org/vocab/InC/1.0/
153-160
oai:omia:id:27678
2015-10-09T01:00:00Z
Classification of SD-OCT Volumes with LBP: Application to DME Detection
Classification of SD-OCT Volumes with LBP: Application to DME Detection
Lemaître, Guillaume
Rastgoo, Mojdeh
Massich, Joan
Sankar, Shrinivasan
Mériaudeau, Fabrice
Sidibé, Désiré
This paper addresses the problem of automatic classification of Spectral Domain OCT (SD-OCT) data for automatic identification of patients with Diabetic Macular Edema (DME) versus normal subjects. Our method is based on Local Binary Patterns (LBP) features to describe the texture of Optical Coherence Tomography (OCT) images and we compare different LBP features extraction approaches to compute a single signature for the whole OCT volume. Experimental results with two datasets of respectively 32 and 30 OCT volumes show that regardless of using low or high level representations, features derived from LBP texture have highly discriminative power. Moreover, the experiments show that the proposed method achieves better classification performances than other recent published works.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1021
https://doi.org/10.17077/omia.1021
https://pubs.lib.uiowa.edu/omia/article/id/27678/
http://rightsstatements.org/vocab/InC/1.0/
9-16
oai:omia:id:27674
2015-10-09T01:00:00Z
Curvature Based Biomarkers for Diabetic Retinopathy via Exponential Curve Fits in SE(2)
Curvature Based Biomarkers for Diabetic Retinopathy via Exponential Curve Fits in SE(2)
Bekkers, Erik J.
Zhang, Jiong
Duits, Remco
ter Haar Romeny, Bart M.
We propose a robust and fully automatic method for the analysis of vessel tortuosity. Our method does not rely on pre-segmentation of vessels, but instead acts directly on retinal image data. The method is based on theory of best-fit exponential curves in the roto-translation group SE(2). We lift 2D images to 3D functions called orientation scores by including an orientation dimension in the domain. In the extended domain of positions and orientations (identified with SE(2)) we study exponential curves, whose spatial projections have constant curvature. By locally fitting such curves to data in orientation scores, via our new iterative stabilizing refinement method, we are able to assign to each location a curvature and confidence value. These values are then used to define global tortuosity measures. The method is validated on synthetic and retinal images. We show that the tortuosity measures can serve as effective biomarkers for diabetes and different stages of diabetic retinopathy.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1034
https://doi.org/10.17077/omia.1034
https://pubs.lib.uiowa.edu/omia/article/id/27674/
http://rightsstatements.org/vocab/InC/1.0/
113-120
oai:omia:id:27667
2015-10-09T01:00:00Z
Effective Drusen Localization for Early AMD Screening using Sparse Multiple Instance Learning
Effective Drusen Localization for Early AMD Screening using Sparse Multiple Instance Learning
Lu, Huiying
Xu, Yanwu
Wong, Damon W. K.
Liu, Jiang
Age-related Macular Degeneration (AMD) is one of the leading causes of blindness. Automatic screening of AMD has attracted much research effort in recent years because it brings benefits to both patients and ophthalmologists. Drusen is an important clinical indicator for AMD in its early stage. Accurately detecting and localizing drusen are important for AMD detection and grading. In this paper, we propose an effective approach to localize drusen in fundus images. This approach trains a drusen classifier from a weakly labeled dataset, i.e., only the existence of drusen is known but not the exact locations or boundaries, by employing Multiple Instance Learning (MIL). Specifically, considering the sparsity of drusen in fundus images, we employ sparse Multiple Instance Learning to obtain better performance compared with classical MIL. Experiments on 350 fundus images with 96 having AMD demonstrates that on the task of AMD detection, multiple instance learning, both classical and sparse versions, achieve comparable performance compared with fully supervised SVM. On the task of drusen localization, sparse MIL outperforms MIL significantly.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1029
https://doi.org/10.17077/omia.1029
https://pubs.lib.uiowa.edu/omia/article/id/27667/
http://rightsstatements.org/vocab/InC/1.0/
73-80
oai:omia:id:27676
2015-10-09T01:00:00Z
Evaluation of Publicly Available Blood Vessel Segmentation Methods for Retinal Images
Evaluation of Publicly Available Blood Vessel Segmentation Methods for Retinal Images
Vostatek, Pavel
Claridge, Ela
Fält, Pauli
Hauta-Kasari, Markku
Uusitalo, Hannu
Lensu, Lasse
Retinal blood vessel structure is an important indicator of disorders related to diseases, which has motivated the development of various image segmentation methods for the blood vessels. In this study, two supervised and two unsupervised retinal blood vessel segmentation methods are quantitatively compared by using five publicly available databases with the ground truth for the vessels. The parameters of each method were optimized for each database with the motivation to achieve good segmentation performance for the comparison and study the importance of proper selection of parameter values. The results show that parameter optimization does not significantly improve the segmentation performance of the methods when the original data is used. However, the methods’ performance for new data differs significantly. Based on the comparison, Soares method as a supervised approach provided the highest overall accuracy and, thus, the best generalisability. Bankhead and Nguyen methods’ performance were close to each other: Bankhead performed better with ARIADB and STARE, whereas Nguyen was better with DRIVE. Sofka method is available only as an executable and its performance matched the others only with ARIADB.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1037
https://doi.org/10.17077/omia.1037
https://pubs.lib.uiowa.edu/omia/article/id/27676/
http://rightsstatements.org/vocab/InC/1.0/
137-144
oai:omia:id:27670
2015-10-09T01:00:00Z
EyeArt + EyePACS: Automated Retinal Image Analysis For Diabetic Retinopathy Screening in a Telemedicine System
EyeArt + EyePACS: Automated Retinal Image Analysis For Diabetic Retinopathy Screening in a Telemedicine System
Bhaskaranand, Malavika
Cuadros, Jorge
Ramachandra, Chaithanya
Bhat, Sandeep
Nittala, Muneeswar G.
Sadda, Srinivas R.
Solanki, Kaushal
Telemedicine frameworks are key to screening the large, ever-growing diabetic population for preventable blindness due to diabetic retinopathy (DR). Integrating fully-automated screening systems in telemedicine frameworks will make DR screening more efficient, cost-effective, reproducible, and accessible. In this paper, we present the integration of EyeArt, an automated DR screening system, into EyePACS, a telemedicine system for DR screening used in diverse screening settings. EyeArt in- corporates novel image processing and analysis algorithms for assessing image gradability; enhancing images based on median filtering; detecting interest regions and localizing lesions based on multi-scale morphological analysis; and DR screening and thus achieves robustness to the large image variability seen in a telemedicine system such as EyePACS. EyeArt is implemented as a scalable, high-throughput cloud-based system to enable large-scale DR screening. We evaluate the safety and performance of EyeArt on a dataset with 434,023 images from 54,324 patient cases obtained from EyePACS. On this dataset, EyeArt’s screening sensitivity is 90% at specificity 60.8% and the area under the receiver operating characteristic curve (AUROC) is 0.883. In a setup where trained human graders review patient cases recommended for referral by EyeArt with low confidence, a workload reduction of 62% is possible. Therefore, EyeArt can be safely integrated into large real world telemedicine DR screening programs such as EyePACS helping reduce workload and increase efficiency and thus help in reducing vision loss due to DR through early detection and treatment.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1033
https://doi.org/10.17077/omia.1033
https://pubs.lib.uiowa.edu/omia/article/id/27670/
http://rightsstatements.org/vocab/InC/1.0/
105-112
oai:omia:id:27675
2015-10-09T01:00:00Z
Geodesic Graph Cut Based Retinal Fluid Segmentation in Optical Coherence Tomography
Geodesic Graph Cut Based Retinal Fluid Segmentation in Optical Coherence Tomography
Bogunović, Hrvoje
Abràmoff, Michael D.
Sonka, Milan
Age-related macular degeneration (AMD) is a leading cause of blindness in developed countries. Its most damaging form is characterized by accumulation of fluid inside the retina, whose quantification is of utmost importance for evaluating the disease progression. In this paper we propose an automated method for retinal fluid segmentation from 3D images acquired with optical coherence tomography (OCT). It combines a machine learning approach with an effective segmentation framework based on geodesic graph cut. After an image preprocessing step, an artificial neural network is trained based on textural features to assign to each voxel a probability of belonging to a fluid. The obtained probability maps are used to compute minimal geodesic distances from a set of identified seed points to the remaining unassigned voxels. Finally, the segmentation is solved optimally and efficiently using graph cut optimization. The method is evaluated on a clinical longitudinal dataset consisting of 30 OCT scans from 10 patients taken at 3 different stages of treatment. Manual annotations from two retinal specialists were taken as the gold standard. The segmentation method achieved mean precision of 0.88 and recall of 0.83, with the combined F1 score of 0.85. The segmented fluid volumes were within the measured inter-observer variability. The results demonstrate that the proposed method is a promising step towards accurate quantification of retinal fluid.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1026
https://doi.org/10.17077/omia.1026
https://pubs.lib.uiowa.edu/omia/article/id/27675/
http://rightsstatements.org/vocab/InC/1.0/
49-56
oai:omia:id:27679
2015-10-09T01:00:00Z
Glaucoma Detection by Learning from Multiple Informatics Domains
Glaucoma Detection by Learning from Multiple Informatics Domains
Xu, Yanwu
Duan, Lixin
Wong, Damon Wing Kee
Wong, Tien Yin
Liu, Jiang
We present a comprehensive and fully automatic glaucoma detection approach that uses machine learning techniques over multiple informatics domains, consisting of personal profile data, genetic data, and retinal image data. This approach, referred to as MKLclm, enriches the feature set of the multiple kernel learning (MKL) framework through the incorporation of classemes, which represent the outputs of multiple class-specific classifiers trained from the data of each informatics domain. We validate our MKLclm framework on a population- based dataset consisting of 2258 subjects, achieving an AUC of 94.9% ± 1.7% and a specificity of 88.5% ± 2.7% at 85% sensitivity, which is significantly better than the current clinical standard of care which uses intraocular pressure (IOP) for glaucoma detection. The experiments also demonstrate that MKLclm outperforms the standard SVM method using data from individual domains, as well as the traditional MKL method, showing that this deeper integration of data from different informatics domains can lead to significant gains in holistic glaucoma diagnosis and screening.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1022
https://doi.org/10.17077/omia.1022
https://pubs.lib.uiowa.edu/omia/article/id/27679/
http://rightsstatements.org/vocab/InC/1.0/
17-24
oai:omia:id:27671
2015-10-09T01:00:00Z
Multimodal Graph-Theoretic Approach for Segmentation of the Internal Limiting Membrane at the Optic Nerve Head
Multimodal Graph-Theoretic Approach for Segmentation of the Internal Limiting Membrane at the Optic Nerve Head
Miri, Mohammad Saleh
Robles, Victor A.
Abràmoff, Michael D.
Kwon, Young H.
Garvin, Mona K.
In this work, we present a multimodal multiresolution graph-based method to segment the top surface of the retina called the internal limiting membrane (ILM) within optic-nerve-head-centered spectral-domain optical coherence tomography (SD-OCT) volumes. Having a precise ILM surface is crucial as this surface is utilized for measuring several structural parameters such as Bruch’s membrane opening-minimum rim width (BMO-MRW) and cup volume. The proposed method addresses the common current segmentation errors due to the presence of retinal blood vessels, deep cupping, or a very steep slope of the ILM. In order to resolve these issues, the volume is resampled using a set of gradient vector flow (GVF) based columns. The GVF field is computed according to an initial surface segmentation which is obtained through a multiresolution framework. The retinal blood vessel information (obtained from corresponding registered fundus photographs) along with shape prior information are incorporated in a graph-theoretic approach to compute the ILM segmentation. The method is tested on the SD-OCT volumes from 44 glaucoma subjects and significantly smaller errors were obtained than that from current approaches.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1027
https://doi.org/10.17077/omia.1027
https://pubs.lib.uiowa.edu/omia/article/id/27671/
http://rightsstatements.org/vocab/InC/1.0/
57-64
oai:omia:id:27666
2015-10-09T01:00:00Z
Obtaining Consensus Annotations For Retinal Image Segmentation Using Random Forest And Graph Cuts
Obtaining Consensus Annotations For Retinal Image Segmentation Using Random Forest And Graph Cuts
Mahapatra, Dwarikanath
Buhmann, Joachim M.
We combine random forest (RF) classifiers and graph cuts (GC) to generate a consensus segmentation of multiple experts. Supervised RFs quantify the consistency of an annotator through a normalized consistency score, while semi supervised RFs predict missing expert annotations. The normalized score is used as the penalty cost in a second order Markov random field (MRF) cost function and the final consensus label is obtained by GC optimization. Experimental results on real patient retinal image datasets show the consensus segmentation by our method is more accurate than those obtained by competing methods.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1025
https://doi.org/10.17077/omia.1025
https://pubs.lib.uiowa.edu/omia/article/id/27666/
http://rightsstatements.org/vocab/InC/1.0/
41-48
oai:omia:id:27668
2015-10-09T01:00:00Z
Refining Coarse Manual Segmentations with Stable Probability Regions
Refining Coarse Manual Segmentations with Stable Probability Regions
Laaksonen, Lauri
Herttuainen, Joni
Uusitalo, Hannu
Lensu, Lasse
Most feature-based lesion detection and computer-aided diagnosis methods for medical images require representative data of each region of interest (ROI) for parameter selection. Furthermore, the spatial accuracy of the segmentation of the ROIs from the background can significantly affect certain image features extracted from the ROIs. How- ever, requiring spatially accurate manual segmentations of the ROIs to be used as the ground truth is infeasible for large image sets due to the amount of manual work involved. To relax the requirement of spatial accuracy and to enable spatial refinement of coarse manual segmentations to have more representative feature data, a method based on color information and maximally stable extremal regions of lesion likelihoods is presented. The proposed method is quantitatively compared to several segmentation approaches by using a challenging set of retinal images with spatially accurate ground truth of exudates. The experiments show that the proposed method produces good results measured as Dice coefficients between the refined segmentation and ground truth.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1031
https://doi.org/10.17077/omia.1031
https://pubs.lib.uiowa.edu/omia/article/id/27668/
http://rightsstatements.org/vocab/InC/1.0/
89-96
oai:omia:id:27680
2015-10-09T01:00:00Z
Retinal Artery/Vein Classification via Graph Cut Optimization
Retinal Artery/Vein Classification via Graph Cut Optimization
Eppenhof, Koen
Bekkers, Erik
Berendschot, Tos T.J.M.
Pluim, Josien P.W.
ter Haar Romeny, Bart M.
In many diseases with a cardiovascular component, the geometry of microvascular blood vessels changes. These changes are specific to arteries and veins, and can be studied in the microvasculature of the retina using retinal photography. To facilitate large-scale studies of artery/vein-specific changes in the retinal vasculature, automated classification of the vessels is required. Here we present a novel method for artery/vein classification based on local and contextual feature analysis of retinal vessels. For each vessel, local information in the form of a transverse intensity profile is extracted. Crossings and bifurcations of vessels provide contextual information. The local and contextual features are integrated into a non-submodular energy function, which is optimized exactly using graph cuts. The method was validated on a ground truth data set of 150 retinal fundus images, achieving an accuracy of 88.0% for all vessels and 94.0% for the six arteries and six veins with highest caliber in the image.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1035
https://doi.org/10.17077/omia.1035
https://pubs.lib.uiowa.edu/omia/article/id/27680/
http://rightsstatements.org/vocab/InC/1.0/
121-128
oai:omia:id:27677
2015-10-09T01:00:00Z
Segmentation of Corneal Endothelial Cells Contour by Means of a Genetic Algorithm
Segmentation of Corneal Endothelial Cells Contour by Means of a Genetic Algorithm
Scarpa, Fabio
Ruggeri, Alfredo
Corneal images acquired by in-vivo microscopy provide clinical information on the cornea endothelium health state. The reliable estimation of the clinical morphometric parameters requires the accurate detection of cell contours in a large number of cells. Thus for the practical application of this analysis in clinical settings an automated method is needed. We propose the automatic segmentation of corneal endothelial cells contour through an innovative technique based on a genetic algorithm, which combines information about the typical regularity of endothelial cells shape with the pixels intensity of the actual image. Ground truth values for the clinical parameters were obtained from manually drawn cell contours. Results show that an accurate automatic estimation is achieved: for each parameter, the mean difference between its manual estimation and the automated one is always less than 4%, and the maximum difference is always less than 7%.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1023
https://doi.org/10.17077/omia.1023
https://pubs.lib.uiowa.edu/omia/article/id/27677/
http://rightsstatements.org/vocab/InC/1.0/
25-32
oai:omia:id:27665
2015-10-09T01:00:00Z
Segmentation of the Retinal Vasculature within Spectral-Domain Optical Coherence Tomography Volumes of Mice
Segmentation of the Retinal Vasculature within Spectral-Domain Optical Coherence Tomography Volumes of Mice
Deng, Wenxiang
Antony, Bhavna
Sohn, Elliott H.
Abràmoff, Michael D.
Garvin, Mona K.
Automated approaches for the segmentation of the retinal vessels are helpful for longitudinal studies of mice using spectral-domain optical coherence tomography (SD-OCT). In the SD-OCT volumes of human eyes, the retinal vasculature can be readily visualized by creating a projected average intensity image in the depth direction. The created projection images can then be segmented using standard approaches. However, in the SD-OCT volumes of mouse eyes, the creation of projection images from the entire volume typically results in very poor images of the vasculature. The purpose of this work is to present and evaluate three machine-learning approaches, namely baseline, single-projection, and all-layers approaches, for the automated segmentation of retinal vessels within SD-OCT volumes of mice. Twenty SD-OCT volumes (400 × 400 × 1024 voxels) from the right eyes of twenty mice were obtained using a Bioptigen SD-OCT machine (Morrisville, NC) to evaluate our methods. The area under the curve (AUC) for the receiver operating characteristic (ROC) curves of the all-layers approach, 0.93, was significantly larger than the AUC for the single-projection (0.91) and baseline (0.88) approach with p < 0.05.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1028
https://doi.org/10.17077/omia.1028
https://pubs.lib.uiowa.edu/omia/article/id/27665/
http://rightsstatements.org/vocab/InC/1.0/
65-72
oai:omia:id:27664
2015-10-09T01:00:00Z
Stability Analysis of Fractal Dimension in Retinal Vasculature
Stability Analysis of Fractal Dimension in Retinal Vasculature
Huang, Fan
Zhang, Jiong
Bekkers, Erik J.
Dashtbozorg, Behdad
ter Haar Romeny, Bart M.
Fractal dimension (FD) has been considered as a potential biomarker for retina-based disease detection. However, conflicting findings can be found in the reported literature regarding the association of the biomarker with diseases. This motivates us to examine the stability of the FD on different (1) vessel segmentations obtained from human observers, (2) automatic segmentation methods, (3) threshold values, and (4) region-of-interests. Our experiments show that the corresponding relative errors with respect to reference ones, computed per patient, are generally higher than the relative standard deviation of the reference values themselves (among all patients). The conclusion of this paper is that we cannot fully rely on the studied FD values, and thus do not recommend their use in quantitative clinical applications.
2015-10-09T01:00:00Z
info:eu-repo/semantics/conferenceObject
2
2015
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1020
https://doi.org/10.17077/omia.1020
https://pubs.lib.uiowa.edu/omia/article/id/27664/
http://rightsstatements.org/vocab/InC/1.0/
1-8
oai:omia:id:27649
2014-09-14T01:00:00Z
ACHIKO-D350: A dataset for early AMD detection and drusen segmentation
ACHIKO-D350: A dataset for early AMD detection and drusen segmentation
Liu, Huiying
Xu, Yanwu
Wong, Damon W.K.
Laude, Augustinus
Lim, Tock Ham
Liu, Jiang
Age related macular degeneration is the third leading cause of global blindness. Its prevalence is increasing in these years for the coming of ”aging population”. Early detection and grading can prevent it from becoming severe and protect vision. Drusen is an important indicator for AMD. Thus automatic drusen detection and segmentation has attracted much research attention in the past years. However, a barrier handicapping the research of drusen segmentation is the lack of a public dataset and test platform. To address this issue, in this paper, we publish a dataset, named ACHIKO-D350, with manually marked drusen boundary. ACHIKO-D350 includes 254 healthy fundus images and 96 fundus images with drusen. The images with drusen cover a wide range of types, including images with sparsely distributed drusen or clumped drusen, images of poor quality, and both well macular centered images and mis-centered images. ACHIKO-D350 will be used for performance evaluation of drusen segmentation methods. It will facilitate an objective evaluation and comparison.
2014-09-14T01:00:00Z
info:eu-repo/semantics/conferenceObject
1
2014
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1011
https://doi.org/10.17077/omia.1011
https://pubs.lib.uiowa.edu/omia/article/id/27649/
http://rightsstatements.org/vocab/InC/1.0/
73-80
oai:omia:id:27662
2014-09-14T01:00:00Z
ACHIKO-M Database for high myopia analysis and its evaluation
ACHIKO-M Database for high myopia analysis and its evaluation
Yin, Fengshou
Li, Ruoying
Zhang, Zhuo
Cheng, Jun
Xu, Yanwu
Wong, Damon Wee Kee
Tan, Ngan Meng
Quan, Ying
Yow, Ai Ping
Kavitha, Gopalakrishnan
Xu, Guozhen
Liu, Jiang
Myopia is the leading public health concern with high prevalence in developed countries. In this paper, we present the ACHIKO-M fundus image database with both myopic and emmetropic cases for high myopia study. The database contains 705 myopic subjects and 151 normal subjects with both left eye and right eye images for each subject. In addition, various clinical data is also available, allowing correlation study of different risk factors. We evaluated two state-of-the-art automated myopia detection algorithms on this database to show how it can be used. Both methods achieve more than 90% accuracy for myopia diagnosis. We will also discuss how ACHIKO-M can be a good database for both scientific and clinical research of myopia.
2014-09-14T01:00:00Z
info:eu-repo/semantics/conferenceObject
1
2014
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1017
https://doi.org/10.17077/omia.1017
https://pubs.lib.uiowa.edu/omia/article/id/27662/
http://rightsstatements.org/vocab/InC/1.0/
121-128
oai:omia:id:27657
2014-09-14T01:00:00Z
An automated framework of inner segment/outer segment defect detection for retinal SD-OCT images
An automated framework of inner segment/outer segment defect detection for retinal SD-OCT images
Zhu, Weifang
Shi, Fei
Xiang, Dehui
Gao, Enting
Wang, Liyun
Chen, Haoyun
Chen, Xinjian
The integrity of inner segment/outer segment (IS/OS) has high correlation with lower visual acuity in patients suffering from blunt trauma. An automated 3D IS/OS defect detection method based on the SD-OCT images was proposed. First, 11 surfaces were automatically segmented using the multiscale 3D graph-search approach. Second, the sub-volumes between surface 7 and 8 containing IS/OS region around the fovea (diameter of mm) were extracted and flattened based on the segmented retinal pigment epithelium layer. Third, 5 kinds of texture based features were extracted for each voxel. A KNN classifier was trained and each voxel was classified as disrupted or nondisrupted and the responding defect volume was calculated. The proposed method was trained and tested on 9 eyes from 9 trauma subjects using the leave-one-out cross validation method. The preliminary results demonstrated the feasibility and efficiency of the proposed method.
2014-09-14T01:00:00Z
info:eu-repo/semantics/conferenceObject
1
2014
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1008
https://doi.org/10.17077/omia.1008
https://pubs.lib.uiowa.edu/omia/article/id/27657/
http://rightsstatements.org/vocab/InC/1.0/
49-56
oai:omia:id:27654
2014-09-14T01:00:00Z
A quadrature filter approach for registration accuracy assessment of fundus images
A quadrature filter approach for registration accuracy assessment of fundus images
Adal, Kedir M.
Couvert, Rosalie
Meijer, D. W. J.
Martinez, Jose P.
Vermeer, Koenraad A.
van Vliet, L. J.
This paper presents a method to automatically assess the accuracy of image registration. It is applicable to images in which vessels are the main landmarks such as fundus images and angiography. The method simultaneously exploits not only the position, but also the intensity profile across the vasculatures. The accuracy measure is defined as the energy of the odd component of the 1D vessel profile in the difference image divided by the total energy of the corresponding vessels in the constituting images. Scale and orientation-selective quadrature filter banks have been employed to analyze the 1D signal profiles. Subsequently, the relative energy measure has been calibrated such that the measure translates to a spatial misalignment in pixels. The method was validated on a fundus image dataset from a diabetic retinopathy screening program at the Rotterdam Eye Hospital. An evaluation showed that the proposed measure assesses the registration accuracy with a bias of -0.1 pixels and a precision (standard deviation) of 0.9 pixels. The small Fourier footprint of the orientation selective quadrature filters makes the method robust against noise.
2014-09-14T01:00:00Z
info:eu-repo/semantics/conferenceObject
1
2014
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1006
https://doi.org/10.17077/omia.1006
https://pubs.lib.uiowa.edu/omia/article/id/27654/
http://rightsstatements.org/vocab/InC/1.0/
33-40
oai:omia:id:27650
2014-09-14T01:00:00Z
Automated detection and classification of nuclei in immunohistochemical stainings for Fuchs' endothelial corneal dystrophy
Automated detection and classification of nuclei in immunohistochemical stainings for Fuchs' endothelial corneal dystrophy
Janssens, Thomas
De Roo, An-Katrien
Foets, Beatrijs
van den Oord, Joost J.
Van den Berghe, Greet
Guiza, Fabian
Fuchs’ endothelial corneal dystrophy (FECD) is a degenerative disease that affects the elderly population, and which lacks a unifying pathogenic theory and tangible drug targets. Immunohistochemical stainings can be used to identify proteins involved in the pathogenesis of FECD. We introduce a method for the automatic quantification of the ratio of stained cells starting from full high-resolution cornea images. First, the endothelium is extracted using entropy information in a low-resolution resampling. Then, within the endothelium, we heuristically detect and classify nuclei based on their size, color, and the color of the surrounding cytoplasm. This method achieves comparable results to manual evaluation in a set of corneas of patients with and without FECD.
2014-09-14T01:00:00Z
info:eu-repo/semantics/conferenceObject
1
2014
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1013
https://doi.org/10.17077/omia.1013
https://pubs.lib.uiowa.edu/omia/article/id/27650/
http://rightsstatements.org/vocab/InC/1.0/
89-96
oai:omia:id:27653
2014-09-14T01:00:00Z
Comparison of image registration methods for composing spectral retinal images
Comparison of image registration methods for composing spectral retinal images
Laaksonen, Lauri
Claridge, Ela
Falt, Pauli
Hauta-Kasari, Markku
Uusitalo, Hannu
Lensu, Lasse
Spectral retinal images have signficant potential for improving the early detection and visualization of subtle changes due to eye diseases and many systemic diseases. High resolution in both the spatial and the spectral domain can be achieved by capturing a set of narrowband channel images from which the spectral images are composed. With imaging techniques where the eye movement between the acquisition of the images is unavoidable, image registration is required. In this paper, the applicability of the state-of-the-art image registration methods for the composition of spectral retinal images is studied. The registration methods are quantitatively compared using synthetic channel image data of an eye phantom and semisynthetic set of retinal channel images subjected to known transformations. The experiments show that Generalized dual-bootstrap iterative closest point method outperforms the other evaluated methods in registration accuracy and the number of successful registrations.
2014-09-14T01:00:00Z
info:eu-repo/semantics/conferenceObject
1
2014
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1009
https://doi.org/10.17077/omia.1009
https://pubs.lib.uiowa.edu/omia/article/id/27653/
http://rightsstatements.org/vocab/InC/1.0/
57-64
oai:omia:id:27645
2014-09-14T01:00:00Z
Comparison of retinal thickness measurements of normal eyes between topcon algorithm and a graph based algorithm
Comparison of retinal thickness measurements of normal eyes between topcon algorithm and a graph based algorithm
Gao, Enting
Shi, Fei
Zhu, Weifang
Chen, Binyao
Chen, Haoyu
Chen, Xinjian
To assess the agreement between Topcon built-in algorithm and our developed graph based algorithm, the retinal thickness of 9-sectors on an Early Treatment of Diabetic Retinopathy Study(ETDRS) chart measurements for normal subjects was compared. A total of fifty eyes were enrolled in this study. The overall and sectoral thickness on ETDRS chart were calculated using Topcon built-in algorithm and our developed three-dimensional graph based algorithm. Correlation analysis and agreement analysis were performed between the commercial algorithm and our algorithm. A high degree of correlation was found between the results obtained from the two methods was from 0.856 to 0.960. It’s showed that our developed graph based algorithm can provide excellent performance similar to Topcon algorithm.
2014-09-14T01:00:00Z
info:eu-repo/semantics/conferenceObject
1
2014
University of Iowa
Proceedings of the Ophthalmic Medical Image Analysis International Workshop
10.17077/omia.1012
https://doi.org/10.17077/omia.1012
https://pubs.lib.uiowa.edu/omia/article/id/27645/
http://rightsstatements.org/vocab/InC/1.0/
81-88
metadataPrefix%3Doai_dc%26page%3D2