R. P. Tornow

Multi-frame Super-resolution with Quality Self-assessment for Retinal Fundus Videos

This paper proposes a novel super-resolution framework to reconstruct high-resolution fundus images from multiple low-resolution video frames in retinal fundus imaging. Natural eye movements during an examination are used as a cue for super-resolution in a robust maximum a-posteriori scheme. In order to compensate heterogeneous illumination on the fundus, we integrate retrospective illumination correction for photometric registration to the underlying imaging model. Our method utilizes quality self-assessment to provide objective quality scores for reconstructed images as well as to select regularization parameters automatically. In our evaluation on real data acquired from six human subjects with a low-cost video camera, the proposed method achieved considerable enhancements of low-resolution frames and improved noise and sharpness characteristics by 74%. In terms of image analysis, we demonstrate the importance of our method for the improvement of automatic blood vessel segmentation as an example application, where the sensitivity was increased by 13% using super-resolution reconstruction.

Analysis of Visual Appearance of Retinal Nerve Fibers in High Resolution Fundus Images: A Study on Normal Subjects

The retinal ganglion axons are an important part of the visual system, which can be directly observed by fundus camera. The layer they form together inside the retina is the retinal nerve fiber layer (RNFL). This paper describes results of a texture RNFL analysis in color fundus photographs and compares these results with quantitative measurement of RNFL thickness obtained from optical coherence tomography on normal subjects. It is shown that local mean value, standard deviation, and Shannon entropy extracted from the green and blue channel of fundus images are correlated with corresponding RNFL thickness. The linear correlation coefficients achieved values 0.694, 0.547, and 0.512 for respective features measured on 439 retinal positions in the peripapillary area from 23 eyes of 15 different normal subjects.

Retinal image registration for eye movement estimation

This paper describes a novel methodology for eye fixation measurement using a unique videoophthalmoscope setup and advanced image registration approach. The representation of the eye movements via Poincare plot is also introduced. The properties, limitations and perspective of this methodology are finally discussed.

Geometry-Based Optic Disk Tracking in Retinal Fundus Videos

Fundus video cameras enable the acquisition of image se- quences to analyze fast temporal changes on the human retina in a non-invasive manner. In this work, we propose a tracking-by-detection scheme for the optic disk to capture the human eye motion on-line dur- ing examination. Our approach exploits the elliptical shape of the optic disk. Therefore, we employ the fast radial symmetry transform for an ef- ficient estimation of the disk center point in successive frames. Large eye movements due to saccades, motion of the head or blinking are detected automatically by a correlation analysis to guide the tracking procedure. In our experiments on real video data acquired by a low-cost video cam- era, the proposed method yields a hit rate of 98% with a normalized median accuracy of 4% of the disk diameter. The achieved frame rate of 28 frames per second enables a real-time application of our approach.

Super-resolved retinal image mosaicing

The acquisition of high-resolution retinal fundus images with a large field of view (FOV) is challenging due to technological, physiological and economic reasons. This paper proposes a fully automatic framework to reconstruct retinal images of high spatial resolution and increased FOV from multiple low-resolution images captured with non-mydriatic, mobile and video-capable but low-cost cameras. Within the scope of one examination, we scan different regions on the retina by exploiting eye motion conducted by a patient guidance. Appropriate views for our mosaicing method are selected based on optic disk tracking to trace eye movements. For each view, one super-resolved image is reconstructed by fusion of multiple video frames. Finally, all super-resolved views are registered to a common reference using a novel polynomial registration scheme and combined by means of image mosaicing. We evaluated our framework for a mobile and low-cost video fundus camera. In our experiments, we reconstructed retinal images of up to 30° FOV from 10 complementary views of 15° FOV. An evaluation of the mosaics by human experts as well as a quantitative comparison to conventional color fundus images encourage the clinical usability of our framework.

Analysis of the Retinal Nerve Fiber Layer Texture Related to the Thickness Measured by Optical Coherence Tomography

The retinal nerve fiber layer (RNFL) is one of the most affected retinal structures due to the glaucoma disease. Progression of this disease results in the RNFL atrophy that can be detected as the decrease of the layer’s thickness. Usually, the RNFL thickness can be assessed by optical coherence tomography (OCT). However, an examination using OCT is rather expensive and still not widely available. On the other hand, fundus camera is considered as a common and fundamental diagnostic device utilized at many ophthalmic facilities worldwide. This contribution presents a novel approach to texture analysis enabling assessment of the RNFL thickness in widely used colour fundus photographs. The aim is to propose a regression model based on different texture features effective for description of changes in the RNFL textural appearance related to the variations of RNFL thickness. The performance evaluation uses OCT as a gold standard modality for validation of the proposed approach. The results show high correlation between the models predicted output and RNFL thickness directly measured by OCT.

Blood vessel segmentation in video-sequences from the human retina

This paper deals with the retinal blood vessel segmentation in fundus video-sequences acquired by experimental fundus video camera. Quality of acquired video-sequences is relatively low and fluctuates across particular frames. Especially, due to the low resolution, poor signal-to-noise ratio, and varying illumination conditions within the frames, application of standard image processing methods might be difficult in such experimental fundus images. In this study, we tried two methods for the segmentation of retinal vessels - matched filtering and Hessian-based approach, originally developed for vessel segmentation in standard fundus images. We showed that modified versions of these two approaches, combined with support vector machine (SVM), can be used also for segmentation in experimental low-quality fundus video-sequences. The SVM classifier trained and consecutively tested on the database of high-resolution images achieved classification accuracy over 94 % and thus revealed a possible applicability of the proposed method on low-quality data. Then, testing on low-quality video-sequences revealed sufficiently large reliability in term of segmentation stability within the sequence with the interframe variability in image quality.