Influence of background preprocessing on the performance of deep learning retinal vessel detection.

Purpose: Segmentation of the vessel tree from retinal fundus images can be used to track changes in the retina and be an important first step in a diagnosis. Manual segmentation is a time-consuming process that is prone to error; effective and reliable automation can alleviate these problems but one of the difficulties is uneven image background, which may affect segmentation performance. Approach: We present a patch-based deep learning framework, based on a modified U-Net architecture, that automatically segments the retinal blood vessels from fundus images. In particular, we evaluate how various pre-processing techniques, images with either no processing, N4 bias field correction, contrast limited adaptive histogram equalization (CLAHE), or a combination of N4 and CLAHE, can compensate for uneven image background and impact final segmentation performance. Results: We achieved competitive results on three publicly available datasets as a benchmark for our comparisons of pre-processing techniques. In addition, we introduce Bayesian statistical testing, which indicates little practical difference ( Pr > 0.99 ) between pre-processing methods apart from the sensitivity metric. In terms of sensitivity and pre-processing, the combination of N4 correction and CLAHE performs better in comparison to unprocessed and N4 pre-processing ( Pr > 0.87 ); but compared to CLAHE alone, the differences are not significant ( Pr ≈ 0.38 to 0.88). Conclusions: We conclude that deep learning is an effective method for retinal vessel segmentation and that CLAHE pre-processing has the greatest positive impact on segmentation performance, with N4 correction helping only in images with extremely inhomogeneous background illumination.

Quantitation of microcomputed tomography-imaged ocular microvasculature.

PURPOSE: To quantitatively assess microvascular dimensions in the eyes of neonatal wild-type and VEGF(120)-tg mice, using a novel combination of techniques which permit three-dimensional (3D) image reconstruction. METHODS: A novel combination of techniques was developed for the accurate 3D imaging of the microvasculature and demonstrated on the hyaloid vasculature of the neonatal mouse eye. Vascular corrosion casting is used to create a stable replica of the vascular network and X-ray microcomputed tomography (muCT) to obtain the 3D images. In-house computer-aided image analysis techniques were then used to perform a quantitative morphological analysis of the images. RESULTS: With the use of these methods, differences in the numbers of vessel segments, their diameter, and volume of vessels in the vitreous compartment were quantitated in wild-type neonatal mice or littermates over-expressing a labile (nonheparin binding) isoform of vascular endothelial growth factor (VEGF(120)) from the developing lens. This methodology was instructive in demonstrating that hyaloid vascular networks in VEGFA(120) over-expressing mice have a 10-fold increase in blind-ended, a six-fold increase in connected vessel segments, in addition to a sixfold increase (0.0314 versus 0.0051 mm(3)) in total vitreous vessel volume compared with wild type. These parameters are not readily quantified via histological, ultrastructural, or stereological analysis. CONCLUSION: The combination of techniques described here provides the first 3D quantitative characterization of vasculature in an organ system; i.e., the neonatal murine intra-ocular vasculature in both wild-type mice and a transgenic model of lens-specific over-expression of VEGF.

Synchrotron X-ray microtomography for assessment of bone tissue scaffolds.

X-ray microtomography (microCT) is a popular tool for imaging scaffolds designed for tissue engineering applications. The ability of synchrotron microCT to monitor tissue response and changes in a bioactive glass scaffold ex vivo were assessed. It was possible to observe the morphology of the bone; soft tissue ingrowth and the calcium distribution within the scaffold. A second aim was to use two newly developed compression rigs, one designed for use inside a laboratory based microCT machine for continual monitoring of the pore structure and crack formation and another designed for use in the synchrotron facility. Both rigs allowed imaging of the failure mechanism while obtaining stress-strain data. Failure mechanisms of the bioactive glass scaffolds were found not to follow classical predictions for the failure of brittle foams. Compression strengths were found to be 4.5-6 MPa while maintaining an interconnected pore network suitable for tissue engineering applications.

Time-Resolved Tomographic Quantification of the Microstructural Evolution of Ice Cream.

Ice cream is a complex multi-phase colloidal soft-solid and its three-dimensional microstructure plays a critical role in determining the oral sensory experience or mouthfeel. Using in-line phase contrast synchrotron X-ray tomography, we capture the rapid evolution of the ice cream microstructure during heat shock conditions in situ and operando, on a time scale of minutes. The further evolution of the ice cream microstructure during storage and abuse was captured using ex situ tomography on a time scale of days. The morphology of the ice crystals and unfrozen matrix during these thermal cycles was quantified as an indicator for the texture and oral sensory perception. Our results reveal that the coarsening is due to both Ostwald ripening and physical agglomeration, enhancing our understanding of the microstructural evolution of ice cream during both manufacturing and storage. The microstructural evolution of this complex material was quantified, providing new insights into the behavior of soft-solids and semi-solids, including many foodstuffs, and invaluable data to both inform and validate models of their processing.

The Bayesian operating point of the Canny edge detector.

We have investigated the operating point of the Canny edge detector which minimizes the Bayes risk of misclassification. By considering each of the sequential stages which constitute the Canny algorithm, we conclude that the linear filtering stage of Canny, without postprocessing, performs very poorly by any standard in pattern recognition and achieves error rates which are almost indistinguishable from a priori classification. We demonstrate that the edge detection performance of the Canny detector is due almost entirely to the postprocessing stages of nonmaximal suppression and hysteresis thresholding.

An improved rotation-invariant thinning algorithm.

Ahmed and Ward have recently presented an elegant, rule-based rotation-invariant thinning algorithm to produce a single-pixel wide skeleton from a binary image. We show examples where this algorithm fails on two-pixel wide lines and propose a modified method which corrects this shortcoming based on graph connectivity.

Performance assessment of feature detection algorithms: a methodology and case study on corner detectors.

In this paper, we describe a generic methodology for evaluating the labeling performance of feature detectors. We describe a method for generating a test set and apply the methodology to the performance assessment of three well-known corner detectors: the Kitchen-Rosenfeld, Paler et al., and Harris-Stephens corner detectors. The labeling deficiencies of each of these detectors is related to their discrimination ability between corners and various of the features which comprise the class of noncorners.

A Generic multi-dimensional feature extraction method using multiobjective genetic programming.

In this paper, we present a generic feature extraction method for pattern classification using multiobjective genetic programming. This not only evolves the (near-)optimal set of mappings from a pattern space to a multi-dimensional decision space, but also simultaneously optimizes the dimensionality of that decision space. The presented framework evolves vector-to-vector feature extractors that maximize class separability. We demonstrate the efficacy of our approach by making statistically-founded comparisons with a wide variety of established classifier paradigms over a range of datasets and find that for most of the pairwise comparisons, our evolutionary method delivers statistically smaller misclassification errors. At very worst, our method displays no statistical difference in a few pairwise comparisons with established classifier/dataset combinations; crucially, none of the misclassification results produced by our method is worse than any comparator classifier. Although principally focused on feature extraction, feature selection is also performed as an implicit side effect; we show that both feature extraction and selection are important to the success of our technique. The presented method has the practical consequence of obviating the need to exhaustively evaluate a large family of conventional classifiers when faced with a new pattern recognition problem in order to attain a good classification accuracy.

A model-based consecutive scanline tracking method for extracting vascular networks from 2-D digital subtraction angiograms.

We propose a new model-based algorithm for the automated tracking of vascular networks in 2-D digital subtraction angiograms. Consecutive scanline profiles are fitted by a parametric imaging model to estimate local vessel center point, radius, edge locations and direction. An adaptive tracking strategy is applied with appropriate termination criteria to track each vessel segment. When tracking stops, to prevent premature termination and to detect bifurcations, a look ahead detection scheme is used to search for possible continuation points of the same vessel segment or those of its bifurcated segments. The proposed algorithm can automatically extract the majority of the vascular network without human interaction other than initializing the start point and direction. Compared to other tracking methods, the proposed method highlights accurate estimation of local vessel geometry. Accurate geometric information and a hierarchical vessel network are obtained which can be used for further quantitative analysis of arterial networks to obtain flow conductance estimates.

Improved sampling of the pareto-front in multiobjective genetic optimizations by steady-state evolution: a pareto converging genetic algorithm.

Previous work on multiobjective genetic algorithms has been focused on preventing genetic drift and the issue of convergence has been given little attention. In this paper, we present a simple steady-state strategy, Pareto Converging Genetic Algorithm (PCGA), which naturally samples the solution space and ensures population advancement towards the Pareto-front. PCGA eliminates the need for sharing/niching and thus minimizes heuristically chosen parameters and procedures. A systematic approach based on histograms of rank is introduced for assessing convergence to the Pareto-front, which, by definition, is unknown in most real search problems. We argue that there is always a certain inheritance of genetic material belonging to a population, and there is unlikely to be any significant gain beyond some point; a stopping criterion where terminating the computation is suggested. For further encouraging diversity and competition, a nonmigrating island model may optionally be used; this approach is particularly suited to many difficult (real-world) problems, which have a tendency to get stuck at (unknown) local minima. Results on three benchmark problems are presented and compared with those of earlier approaches. PCGA is found to produce diverse sampling of the Pareto-front without niching and with significantly less computational effort.