Modular deep neural networks for automatic quality control of retinal optical coherence tomography scans.

Retinal optical coherence tomography (OCT) with intraretinal layer segmentation is increasingly used not only in ophthalmology but also for neurological diseases such as multiple sclerosis (MS). Signal quality influences segmentation results, and high-quality OCT images are needed for accurate segmentation and quantification of subtle intraretinal layer changes. Among others, OCT image quality depends on the ability to focus, patient compliance and operator skills. Current criteria for OCT quality define acceptable image quality, but depend on manual rating by experienced graders and are time consuming and subjective. In this paper, we propose and validate a standardized, grader-independent, real-time feedback system for automatic quality assessment of retinal OCT images. We defined image quality criteria for scan centering, signal quality and image completeness based on published quality criteria and typical artifacts identified by experienced graders when inspecting OCT images. We then trained modular neural networks on OCT data with manual quality grading to analyze image quality features. Quality analysis by a combination of these trained networks generates a comprehensive quality report containing quantitative results. We validated the approach against quality assessment according to the OSCAR-IB criteria by an experienced grader. Here, 100 OCT files with volume, circular and radial scans, centered on optic nerve head and macula, were analyzed and classified. A specificity of 0.96, a sensitivity of 0.97 and an accuracy of 0.97 as well as a Matthews correlation coefficient of 0.93 indicate a high rate of correct classification. Our method shows promising results in comparison to manual OCT grading and may be useful for real-time image quality analysis or analysis of large data sets, supporting standardized application of image quality criteria.

Optic nerve head three-dimensional shape analysis.

We present a method for optic nerve head (ONH) 3-D shape analysis from retinal optical coherence tomography (OCT). The possibility to noninvasively acquire in vivo high-resolution 3-D volumes of the ONH using spectral domain OCT drives the need to develop tools that quantify the shape of this structure and extract information for clinical applications. The presented method automatically generates a 3-D ONH model and then allows the computation of several 3-D parameters describing the ONH. The method starts with a high-resolution OCT volume scan as input. From this scan, the model-defining inner limiting membrane (ILM) as inner surface and the retinal pigment epithelium as outer surface are segmented, and the Bruch's membrane opening (BMO) as the model origin is detected. Based on the generated ONH model by triangulated 3-D surface reconstruction, different parameters (areas, volumes, annular surface ring, minimum distances) of different ONH regions can then be computed. Additionally, the bending energy (roughness) in the BMO region on the ILM surface and 3-D BMO-MRW surface area are computed. We show that our method is reliable and robust across a large variety of ONH topologies (specific to this structure) and present a first clinical application.

Active contour method for ILM segmentation in ONH volume scans in retinal OCT.

The optic nerve head (ONH) is affected by many neurodegenerative and autoimmune inflammatory conditions. Optical coherence tomography can acquire high-resolution 3D ONH scans. However, the ONH's complex anatomy and pathology make image segmentation challenging. This paper proposes a robust approach to segment the inner limiting membrane (ILM) in ONH volume scans based on an active contour method of Chan-Vese type, which can work in challenging topological structures. A local intensity fitting energy is added in order to handle very inhomogeneous image intensities. A suitable boundary potential is introduced to avoid structures belonging to outer retinal layers being detected as part of the segmentation. The average intensities in the inner and outer region are then rescaled locally to account for different brightness values occurring among the ONH center. The appropriate values for the parameters used in the complex computational model are found using an optimization based on the differential evolution algorithm. The evaluation of results showed that the proposed framework significantly improved segmentation results compared to the commercial solution.

Return radius and volume of recrystallized material in Ostwald ripening.

Within the framework of the Lifshitz-Slyozov-Wagner theory of Ostwald ripening, the amount of volume of the second (solid) phase in a liquid solution that is newly formed by recrystallization is investigated. It is shown that in the late stage, the portion of the newly generated volume formed within an interval from time t(0) to t is a certain function of t/t(0) and an explicit expression of this volume is given. To achieve this, we introduce the notion of the return radius r(t,t(0)), which is the unique radius of a particle at time t(0) such that this particle has-after growing and shrinking-the same radius at time t. We derive a formula for the return radius, which later on is used to obtain the newly formed volume. Moreover, formulas for the growth rate of the return radius and the recrystallized material at time t(0) are derived.

Geometric Ginzburg-Landau theory for faceted crystals in one dimension: From coarsening to chaos through a driving force.

We consider the dynamic behavior in driven phase transitions dominated either by attachment-detachment or by surface diffusion mass transport mechanisms. As the driving force increases, we numerically demonstrate for both cases that the spatiotemporal faceted structure of the surface undergoes a sequential transition from slow coarsening turning to accelerated coarsening followed by fixed length scale structures before finally becoming spatiotemporally chaotic. For the attachment-detachment dominated phase transition problem we compare in the accelerated coarsening regime the simulation results with an intrinsic dynamical system governing the leading-order piecewise-affine dynamic surface (PADS), which can be obtained through a matched asymptotic analysis. The PADS predicts the numerically observed coarsening law for the growth in time of the characteristic morphological length scale L_{M} . In particular we determine the prefactor of the scaling law which allows for quantitative predictions necessary for any use of the theory in preparing patterned surfaces through modifications of the driving force.

Morphological stability of electromigration-driven vacancy islands.

The electromigration-induced shape evolution of two-dimensional vacancy islands on a crystal surface is studied using a continuum approach. We consider the regime where mass transport is restricted to terrace diffusion in the interior of the island. In the limit of fast attachment (detachment) kinetics a circle translating at constant velocity is a stationary solution of the problem. In contrast to earlier work [O. Pierre-Louis and T. L. Einstein, Phys. Rev. B 62, 13697 (2000)], we show that the circular solution remains linearly stable for arbitrarily large driving forces. The numerical solution of the full nonlinear problem nevertheless reveals a fingering instability at the trailing end of the island, which develops from finite amplitude perturbations and eventually leads to pinch off. Relaxing the condition of instantaneous attachment (detachment) kinetics, we obtain noncircular elongated stationary shapes in an analytic approximation that compares favorably to the full numerical solution.

Complex shape evolution of electromigration-driven single-layer islands.

The shape evolution of two-dimensional islands through periphery diffusion biased by an electromigration force is studied numerically using a continuum approach. We show that the introduction of crystal anisotropy in the mobility of edge atoms induces a rich variety of migration modes, which include oscillatory and irregular behavior. A phase diagram in the plane of anisotropy strength and island size is constructed. The oscillatory motion can be understood in terms of stable facets that develop on one side of the island, and which the island then slides past. The facet orientations are determined analytically.