Albert M. Vossepoel

A model based method for retinal blood vessel detection

Retinal blood vessels are important structures in ophthalmological images. Many detection methods are available, but the results are not always satisfactory. In this paper, we present a novel model based method for blood vessel detection in retinal images. It is based on a Laplace and thresholding segmentation step, followed by a classification step to improve performance. The last step assures incorporation of the inner part of large vessels with specular reflection. The method gives a sensitivity of 92% with a specificity of 91%. The method can be optimized for the specific properties of the blood vessels in the image and it allows for detection of vessels that appear to be split due to specular reflection.

Adaptive Vectorization of Line Drawing Images

A novel method for vectorizing line drawing images is presented. The method is based on a sequence of a standard vectorization algorithm and maximum threshold morphology, which can be iterated until a fitting criterion is met. First, “anchor points” are found in a coarse vectorization. The position of each anchor point is corrected by morphological operations, after which it is fixed. Next, the vectorization is refined between the anchor points. The method is adaptive because the original input image is used for correcting and improving the vectorization. Postprocessing modules can be added to anticipate specific properties of the line drawings to be vectorized. The vectorization method is evaluated using line drawings differing in scale and resolution.

Automatic classification of ships from infrared (FLIR) images

The aim of the research presented in this paper is to find out whether automatic classification of ships from Forward Looking InfraRed images is feasible in maritime patrol aircraft. An image processing system has been developed for this task. It includes iterative shading correction and a top hat filter for the detection of the ship. It uses a segmentation algorithm based on the gray value distribution of the waves and the Hough transform to locate the waterline of the ship. A model has been developed to relate the size of the ship and the angle between waterline and horizon in image coordinates, to the real-life size and aspect angle of the ship. The model uses the camera elevation and distance to the ship. A data set was used consisting of two civil ships and four different frigates under different aspect angles and distances. From each of these ship images, 32 features were calculated, among which are the apparent size, the location of the hot spot and of the superstructures of the ship, and moment invariant functions. All features were used in feature selection processing using both the Mahalanobis and nearest neighbor (NN) criteria to forward, backward, and branch and bound feature selection procedures, to find the most significant features. Classification has been performed using a k-NN, a linear and quadratic classifier. In particular, using the 1-NN classifier, good results were achieved using a two-step classification algorithm.

FULLY AUTOMATED WHOLE-BODY REGISTRATION IN MICE USING AN ARTICULATED SKELETON ATLAS

In this paper, we propose a fully automated articulated registration approach for whole-body 3D data of mice. The method is based on a hierarchical anatomical model of the skeletal system where we specified position and degrees of freedom for each joint. Model fitting is performed by traversing a hierarchical part-tree, which enables a coarse-to-fine registration from the inner articulations outwards. The method was tested on 12 micro-CT volumes, giving accurate alignment of the skeletal structures in all cases.

Improving Triangle Mesh Quality with SurfaceNets

Simulation of soft tissue deformation is a critical part of surgical simulation. An important method for this is finite element (FE) analysis. Models for FE analysis are typically derived by extraction of triangular surface meshes from CT or MRI image data. These meshes must fulfill requirements of accuracy, smoothness, compactness, and triangle quality. In this paper we propose new techniques for improving mesh triangle quality, based on the SurfaceNets method. Our results show that the meshes created are smooth and accurate, have good triangle quality, and fine detail is retained.

Feature-based detection of landmines in infrared images

High detection performance is required for an operational system for the detection of landmines. Humanitarian de-mining scenarios, combined with inherent difficulties of detecting landmines on an operational (vibration, motion, atmosphere) as well as a scenario level (clutter, soil type, terrain), result in high levels of false alarms for most sensors. To distinguish a landmine from background clutter one or more discriminating object features have to be found. The research described here focuses on finding and evaluating one or more features to distinguish disk-shaped landmines from background clutter in infrared images. These images were taken under controlled conditions, with homogenous soil types. Two methods are considered to acquire shape-based features in the infrared imagery. The first method uses a variation of the Hough transformation to find circular shaped objects. The second method uses the tophat filter with a disk-shaped structuring element. Furthermore, Mahalanobis and Fisher based classifiers are used to combine these features.

A statistical shape model without using landmarks

This paper describes the construction of a statistical shape model based on the iterative closest point algorithm. The method does not require manual nor automatic identification of explicit landmarks on example shapes. Corresponding features are found by retrieving the nearest points via interpolation along the surface. The application to analyse carpal bone shape renders evidence that the lunate bone occurs in distinct types.

Modeling of scanning laser polarimetry images of the human retina for progression detection of glaucoma

The development of methods to detect slowly progressing diseases is often hampered by the time-consuming acquisition of a sufficiently large data set. In this paper, a method is presented to model the change in images acquired by scanning laser polarimetry, for the detection of glaucomatous progression. The model is based on image series of 23 healthy eyes and incorporates colored noise, incomplete cornea compensation and masking by the retinal blood vessels. Additionally, two methods for detecting progression, taking either one or two follow-up visits into account, are discussed and tested on these simulated images. Both methods are based on Students t-tests, morphological operations and anisotropic filtering. The images simulated by the model are visually pleasing, show corresponding statistical properties to the real images and are used to optimize the detection methods. The results show that detecting progression based on two follow-up visits greatly improves the sensitivity without adversely affecting the specificity.

Roentgen Stereophotogrammetric Analysis: An Accurate Tool to Assess Stent-Graft Migration

Purpose: To evaluate in an in vitro model the feasibility and accuracy of Roentgen stereophotogrammetric analysis (RSA) versus computed tomography (CT) for the ability to detect stent-graft migration. Methods: An aortic model was constructed from a 22-mm-diameter Plexiglas tube with 6-mm polytetrafluoroethylene inlays to mimic the renal arteries. Six tantalum markers were placed in the wall of the aortic tube proximal to the renal arteries. Another 6 markers were added to a Gianturco stent, which was cast in Plexiglas and placed inside the aorta and fixed to a micromanipulator to precisely control displacement of the stent along the longitudinal axis. Sixteen migrations were analyzed with RSA software and compared to the micromanipulator. Thirty-two migrations were measured by 3 observers from CT images acquired with 16X0.5-mm beam collimation and reconstructed with a 0.5-mm slice thickness and a 0.4-mm reconstruction interval. Measurements were made with Vitrea postprocessing software using a standard clinical protocol and central lumen line reconstruction. Results of CT were also compared to the micromanipulator. Results: The mean RSA measurement error compared to the micromanipulator was 0.002±0.044 mm, and the maximum error was 0.10 mm. There was no statistically significant interobserver variability for CT (p=0.17). The pooled mean (maximum) measurement error of CT was 0.14±0.29 (1.00) mm, which was significantly different from the RSA measurement error (p<0.0001). Conclusion: Detection of endograft migration by RSA is feasible and was significantly more accurate than CT in this nonpulsatile in vitro model.

Improving the imaging of calcifications in CT by histogram-based selective deblurring

Imaging of small high-density structures, such as calcifications, with computed tomography (CT) is limited by the spatial resolution of the system. Blur causes small calcifications to be imaged with lower contrast and overestimated volume, thereby hampering the analysis of vessels. The aim of this work is to reduce the blur of calcifications by applying three-dimensional (3D) deconvolution. Unfortunately, the high-frequency amplification of the deconvolution produces edge-related ring artifacts and enhances noise and original artifacts, which degrades the imaging of low-density structures. A method, referred to as Histogram-based Selective Deblurring (HiSD), was implemented to avoid these negative effects. HiSD uses the histogram information to generate a restored image in which the low-intensity voxel information of the observed image is combined with the high-intensity voxel information of the deconvolved image. To evaluate HiSD we scanned four in-vitro atherosclerotic plaques of carotid arteries with a multislice spiral CT and with a microfocus CT (μCT), used as reference. Restored images were generated from the observed images, and qualitatively and quantitatively compared with their corresponding μCT images. Transverse views and maximum-intensity projections of restored images show the decrease of blur of the calcifications in 3D. Measurements of the areas of 27 calcifications and total volumes of calcification of 4 plaques show that the overestimation of calcification was smaller for restored images (mean-error: 90% for area; 92% for volume) than for observed images (143%; 213%, respectively). The qualitative and quantitative analyses show that the imaging of calcifications in CT can be improved considerably by applying HiSD.