A Fast Lightweight Based Deep Fusion Learning for Detecting Macula Fovea Using Ultra-Widefield Fundus Images

Macula fovea detection is a crucial prerequisite towards screening and diagnosing macular diseases. Without early detection and proper treatment, any abnormality involving the macula may lead to blindness. However, with the ophthalmologist shortage and time-consuming artificial evaluation, neither accuracy nor effectiveness of the diagnose process could be guaranteed. In this project, we proposed a deep learning approach on ultra-widefield fundus (UWF) images for macula fovea detection. This study collected 2300 ultra-widefield fundus images from Shenzhen Aier Eye Hospital in China. Methods based on U-shape network (Unet) and Fully Convolutional Networks (FCN) are implemented on 1800 (before amplifying process) training fundus images, 400 (before amplifying process) validation images and 100 test images. Three professional ophthalmologists were invited to mark the fovea. A method from the anatomy perspective is investigated. This approach is derived from the spatial relationship between macula fovea and optic disc center in UWF. A set of parameters of this method is set based on the experience of ophthalmologists and verified to be effective. Results are measured by calculating the Euclidean distance between proposed approaches and the accurate grounded standard, which is detected by Ultra-widefield swept-source optical coherence tomograph (UWF-OCT) approach. Through a comparation of proposed methods, we conclude that, deep learning approach of Unet outperformed other methods on macula fovea detection tasks, by which outcomes obtained are comparable to grounded standard method.

A Fast Lightweight Based Deep Fusion Learning for Detecting Macula Fovea Using Ultra-Widefield Fundus Images

Macula fovea detection is a crucial prerequisite towards screening and diagnosing macular diseases. Without early detection and proper treatment, any abnormality involving the macula may lead to blindness. However, with the ophthalmologist shortage and time-consuming artificial evaluation, neither accuracy nor effectiveness of the diagnose process could be guaranteed. In this project, we proposed a deep learning approach on ultra-widefield fundus (UWF) images for macula fovea detection. This study collected 2300 ultra-widefield fundus images from Shenzhen Aier Eye Hospital in China. Methods based on U-shape network (Unet) and Fully Convolutional Networks (FCN) are implemented on 1800 (before amplifying process) training fundus images, 400 (before amplifying process) validation images and 100 test images. Three professional ophthalmologists were invited to mark the fovea. A method from the anatomy perspective is investigated. This approach is derived from the spatial relationship between macula fovea and optic disc center in UWF. A set of parameters of this method is set based on the experience of ophthalmologists and verified to be effective. Results are measured by calculating the Euclidean distance between proposed approaches and the accurate grounded standard, which is detected by Ultra-widefield swept-source optical coherence tomograph (UWF-OCT) approach. Through a comparation of proposed methods, we conclude that, deep learning approach of Unet outperformed other methods on macula fovea detection tasks, by which outcomes obtained are comparable to grounded standard method.

Morphological Study of Color Doppler Ultrasound Imaging to Observe the Effect of Ocular Fundus Surgery

High myopia continues to progress and the eye axis continues to grow, resulting in mechanical dilatation of the eyeball wall, and with the increase of age, resulting in a variety of myopic fundus pathological changes. Considering that in the late stage of fundus disease, patients can obviously feel the impact of the disease on vision, and ophthalmologists are needed for direct diagnosis, intervention and treatment. The impact of early lesions on vision is relatively weak, and patients can not detect the emergence of these lesions in time, which can only be found by fundus survey. Therefore, the automatic detection of early lesions of fundus disease (especially microaneurysms) is not only of great significance for early diagnosis and early treatment of fundus disease, but also a good reference for the detection of other fundus diseases. In this paper, color ultrasound can effectively analyze the origin of fundus diseases. The results showed that with the changes of age, eye axis and posterior sclera shape in high myopia, the area and range of myopic arc and choroidal atrophy arc expanded, the thickness of nerve fiber and choroid became thinner, and the fovea of macula mainly shifted vertically. In addition, it was also found that myopic arc, choroidal atrophic arc area, retinal thickness, nerve fiber thickness, choroidal thickness, vertical distance from macular fovea to optic disc center were correlated with age and eye axis.

Measuring ocular torsion and its variations using different nonmydriatic fundus photographic methods

Purpose To compare the variations in ocular torsion measurements made using different fundus photographic methods. Methods We enrolled subjects with three conditions: (1) patients with intermittent exotropia (IXT) (n = 44), (2) patients with unilateral superior oblique palsy (SOP) (n = 10), and (3) normal subjects as controls (n = 85). Ocular torsion was measured by disc-center–fovea angle (DFA) using three different imaging modalities: (1) conventional fundus photography (CFP) with a 45° field of view (FV), (2) wide-field fundus photography (WFP) with a 200° FV, and (3) optical coherence tomography (OCT) with a 55° FV. Results In the IXT group, the DFAs in the right and left eyes were 5.70±3.35° and 6.37±3.36°, respectively, for CFP, 8.39±5.24° and 8.61±3.67° for WFP, and 5.73±3.61° for 6.16±3.50° for OCT. In the SOP group, the DFAs in paretic and nonparetic eyes were 12.19±1.69° and 6.71±1.09°, respectively, for CFP, 14.29±2.36° and 8.23±3.31° for WFP, and 12.12±1.73° and 6.91±1.12° for OCT. In the control group, the DFAs in the right and left eyes were 5.39±2.65° and 5.71±3.16°, respectively, for CFP, 8.77±5.56° and 8.90±6.24° for WFP, and 5.27±2.67° and 5.72±3.20° for OCT. There was no difference between the results from CFP and OCT among the three groups. However, the torsional angle was larger when measured using WFP than the other two photographic methods (CFP and OCT) in all three groups (all p<0.05). Conclusion The ocular torsion measurement varies with the fundus photographic method used to measure it. Clinicians should be careful to avoid overestimating ocular extorsion when it is evaluated using WFP.

Measurements of spatially unresolved magnetic fields of mixed polarity using observations of the Zeeman effect in the solar faculae

It is shown that it is quite possible to measure spatially unresolved magnetic fields of mixed polarity by the Zeeman effect if their intensity exceeds 150-200 G. This means that one can interconnect the ranges of the registration of these fields based on the Hanle and Zeeman effects. This thesis is illustrated by the results of calculations of the half-width of the Stokes profile I for the FeI 5247.1 and 5250.2 lines. Parameter I is sensitive to the presence of magnetic fields, regardless of whether they have the same magnetic polarity within the input aperture of the instrument, or the opposite one. According to the calculations, if the accuracy of the measurement of the half-width ratio of these two lines is increased to 0.5 %, then it is possible to measure the magnetic fields from 70-100 G. On the basis of the proposed method, an estimate of the intensity of the magnetic fields of mixed polarity in a solar faculae in the tail part of the active region NOAA 1809 was made, which on the day of observations, August 6, 2013, was located not far from the Sun’s disc center. The Echelle Zeeman-spectrogram of this region was obtained on the Horizontal Solar Telescope of Astronomical Observatory of Taras Shevchenko National University of Kyiv. The magnetic field in the faculae was measured in two ways: by shift of of the “center of gravity”of the profiles of the FeI 5247.1 and FeI 5250.2 Ǻ lines in the I + V i I – V spectra and also by the half-width ratio of the profiles I of these lines. The first method allows to measure the effective magnetic field Beff, which turned out to be 280 G by FeI 5250.2 and 360 G by FeI 5247.1. The corresponding ratio Beff (5247.1) / Beff (5250.2) ≈ 1.3 indicates the existence of the sub-telescopic flux tubes with kilogauss fields. However, the ratio of the half-width of the Stokes I profiles of these two lines is 1.08, which corresponds to the magnetic field ± 650 G, if magnetic field is purely longitudinal. Based on the analysis of these data as well as the results of other studies, one can conclude that in the investigated faculae there were probably three magnetic field field components: the fluxtubes with kG field Bfluxtube, the areas of background field Bbackgr of regular magnetic polarity, and the areas of subtelescopic fields Bmixpol of mixed magnetic polarity. In absolute value the magnetic flux of tangled mixed-polarity field exceeds the flux of entire regular field with the intensities of Bfluxtube and Bbackgr at least 2-fold.

Contrast based circular approximation for accurate and robust optic disc segmentation in retinal images

A new method for automatic optic disc localization and segmentation is presented. The localization procedure combines vascular and brightness information to provide the best estimate of the optic disc center which is the starting point for the segmentation algorithm. A detection rate of 99.58% and 100% was achieved for the Messidor and ONHSD databases, respectively. A simple circular approximation to the optic disc boundary is proposed based on the maximum average contrast between the inner and outer ring of a circle centered on the estimated location. An average overlap coefficient of 0.890 and 0.865 was achieved for the same datasets, outperforming other state of the art methods. The results obtained confirm the advantages of using a simple circular model under non-ideal conditions as opposed to more complex deformable models.

Automatic Localization of the Optic Disc Center in Retinal Images Based on Angle Detection in Curvature Scale Space

Digital images of the retina is widely used for screening of patients suffering from sight threatening diseases such as Diabetic retinopathy and Glaucoma. The localization of the Optic Disc (OD) center is the first and necessary step identification and segmentation of anatomical structures and in pathological retinal images. From the center of the optic disc spreads the major blood vessels of the retina. Therefore, by considering the high number of vessels and the high number of the angles resulted from the vessels crossing, the authors propose a new method based on the number of angles in the vicinity of optic disc for localization of the center of optic disc. The first step is pre-processing of retinal image for separate the fundus from its background and increase the contrast between contours. In the second step, the authors use the Curvature Scale Space (CSS) for angle detection. In the next step, they move a window about the size of optic disc to count the number of corners. In the final step, they use the center of windows which has the most number of corners for localizing the optic disc center. The proposed method is evaluated on DRIVE, CHASE_DB1 and STARE databases and the success rate is 100, 100 and 96.3%, respectively.