Accurate object localization in gray level images using the center of gravity measure: accuracy versus precision.

A widely used subpixel precision estimate of an object center is the weighted center of gravity (COG). We derive three maximum-likelihood estimators for the variance of the two-dimensional (2-D) COG as a function of the noise in the image. We assume that the noise is additive, Gaussian distributed and independent between neighboring pixels. Repeated experiments using 2500 generated 2-D bell-shaped markers superimposed with an increasing amount of Gaussian noise were performed, to compare the three approximations. The error of the most exact approximative variance estimate with respect to true variance was always less than 5% of the latter. This deviation decreases with increasing signal-to-noise ratio. Our second approximation to the variance estimate performed better than the third approximation, which was originally presented by Oron et al. by up to a factor approximately 10. The difference in performance between these two approximations increased with an increasing misplacement of the window in which the COG was calculated with respect to the real COG.

Contrast-enhanced magnetic resonance imaging of bone marrow.

In the assessment with magnetic resonance (MR) imaging of bone marrow disorders, the use of contrast agents is usually not critical because T1-weighted spin-echo and fat-suppressed sequences (STIR or fat-sat intermediate weighted) are robust and largely available techniques for depiction of neoplastic and non-neoplastic lesions of the bone marrow. This article discusses the characteristics of dynamic contrast-enhanced MR imaging of bone marrow edema, ischemia, and neoplasm. It emphasizes its value in staging and in monitoring of response to chemotherapy of several bone tumors. These fast dynamic contrast-enhanced techniques do not allow differentiation between benign and malignant primary osseous tumors because the biologic behavior rather than the malignant potential of these lesions is reflected.

Detection of leukocytes in contact with the vessel wall from in vivo microscope recordings using a neural network.

Leukocytes play an important role in the host defense as they may travel from the blood stream into the tissue in reacting to inflammatory stimuli. The leukocyte-vessel wall interactions are studied in post capillary vessels by intravital video microscopy during in vivo animal experiments. Sequences of video images are obtained and digitized with a frame grabber. A method for automatic detection and characterization of leukocytes in the video images is developed. Individual leukocytes are detected using a neural network that is trained with synthetic leukocyte images generated using a novel stochastic model. This model makes it feasible to generate images of leukocytes with different shapes and sizes under various lighting conditions. Experiments indicate that neural networks trained with the synthetic leukocyte images perform better than networks trained with images of manually detected leukocytes. The best performing neural network trained with synthetic leukocyte images resulted in an 18% larger area under the ROC curve than the best performing neural network trained with manually detected leukocytes.

Detection of areas with viable remnant tumor in postchemotherapy patients with Ewing's sarcoma by dynamic contrast-enhanced MRI using pharmacokinetic modeling.

An approach is presented for monitoring the effects of neoadjuvant chemotherapy in patients with Ewing's sarcoma using dynamic contrast-enhanced perfusion magnetic resonance (MR) images. For that purpose, we modify the three-compartment pharmacokinetic permeability model introduced by Tofts et al. (Magn Reson Med 1991;17:357-67) to a two-compartment model. Perfusion MR images acquired using an intravenous injection with Gadolinium (Gd-DTPA) are analyzed with this two-compartment pharmacokinetic model as well as the with an extended pharmacokinetic model that includes the (local) arrival time t(0) of the tracer as an endogenous (estimated) parameter. For each MR section, a wash-in parameter associated with each voxel is estimated twice by fitting each of the two pharmacokinetic models to the dynamic MR signal. A comparison of the two wash-in parametric images (global versus local arrival time) with matched histologic macroslices demonstrates a good correspondence between areas with viable remnant tumor and a high wash-in rate. This can be explained by the high number and permeability of the (leaking) capillaries in viable tumor tissue. The novel pharmacokinetic model based on a local arrival time of tracer results in the best fit of the wash-in rate, the most important factor discerning viable from nonviable tumor components. However, parameter estimates obtained with this model are also more sensitive to noise in the MR signal. The novel pharmacokinetic model resulted in a sensitivity between 0.22 and 0.60 and a specificity between 0.61 and 1. The model based on a global arrival time gave sensitivities between 0.33 and 0.77 and specificities between 0.58 and 0.99. Both statistics are computed as the fraction of correctly labeled voxels (viable or nonviable tumor) within a specified ROI, which delineates the tumor. We conclude that the added value of estimating the local arrival time of tracer first manifests itself for moderate noise levels in the MR signal. The novel pharmacokinetic model should moreover be preferred when pharmacokinetic modeling is applied on the average signal intensity within a ROI, where noise has less effect on the fitted parameters.

Automated calibration in vascular X-ray images using the accurate localization of catheter marker bands.

RATIONALE AND OBJECTIVES: To develop a new automated calibration method for vessel measurements in vascular x-ray images. METHODS: Radiopaque marker bands mounted equidistantly on a small catheter were acquired in vitro at five image intensifier (II) sizes in x-ray projection images. The positions of the marker centers were detected by using a Hough transform and were computed at subpixel precision by using either a novel, iterative center-of-gravity approach (CGA) or a symmetry filter. Curve-fitting procedures were used to reject false-positive marker detections and to calculate intermarker distances. The calibration factor was calculated from the true marker distance and the average of the measured distances in pixels. Results were compared statistically with a grid calibration method, which was taken as the gold standard. A simulation study was performed to assess the influence of image noise on the CGA method. RESULTS: The iterative CGA method was convergent and faster than the symmetry-based technique. For four II sizes (17, 20, 25, and 31 cm), the results from the CGA method were not significantly different from the results obtained with grid calibration. For the II size of 38 cm, a significant difference (0.3% of the grid calibration factor) was found; however, this was caused by the quantification error in the image data and was not clinically relevant. In general, the performance of the CGA method improved with increasing signal-to-noise ratio. CONCLUSIONS: A practical new calibration method for small catheter sizes was developed and validated for quantitative vascular arteriography.

Robustness metrics for measuring the influence of additive noise on the performance of statistical classifiers.

This paper presents a novel quality measure called robustness. The robustness measure quantifies the influence of measurement noise in the attribute values on the credibility of the classification of a case. It is assumed that the type of distribution of the noise-generating process is known. It is not simple to measure the robustness in the general situation where the noise-free distribution of the attributes is unknown. Therefore, two approximations are suggested and compared with the robustness measure based on the noise-free distribution of the attributes. The usefulness of the suggested robustness measure is explored in a simulation experiment.

User-requirements driven learning.

This paper describes an approach for deriving classification knowledge from databases, taking into account user preferences. These preferences especially concern the trade-off between different kinds of costs and performance indicators of the classification scheme to be developed. We analyze what knowledge, provided by the user, can be used at various stages of the machine learning process to influences the development of the classifier. We restrict ourselves in this paper mainly to the generation of classification trees.

On the quality of neural net classifiers.

This paper describes several concepts and metrics that may be used to assess various aspects of the quality of neural net classifiers. Each concept describes a property that may be taken into account by both designers and users of neural net classifiers when assessing their utility. Besides metrics for assessment of the correctness of classifiers we also introduce metrics that address certain aspects of the misclassifications. We show the applicability of the introduced quality concepts for selection among several neural net classifiers in the domain of thyroid disorders.