J. Barba

A parametric fitting algorithm for segmentation of cell images

This paper presents a parametric fitting algorithm for segmentation of cervical and breast cell images from cytology smears. A parametric elliptical model for cells is introduced and the parameters adjusted to fit the cell shapes while minimizing a cost function. Segmentation results of noisy human cervical cell and textured breast cell images demonstrate that the proposed parametric fitting algorithm is very successful in segmentation of images of both nonoverlapped and overlapped elliptically shaped cells.

Ovarian dysplasia: Nuclear texture analysis

Background. Ovarian dysplasia has been defined by histologic1,2 and morphometric studies3,4 focusing on architectural and nuclear profile changes. A new technique is used to enhance the accuracy of this diagnosis by a quantitative evaluation of the nuclear texture that represents the nuclear chromatin pattern on which conventional diagnoses of malignancy are usually made.

An efficient semi‐automatic algorithm for cell contour extraction

An interactive semi‐automatic procedure for extraction of cell contours from light microscope images is presented. The user is required to specify four contour points and the algorithm determines the rest of the contour automatically. The algorithm exploits the fact that cell contours have lower grey level than their immediate surrounding and are usually very similar in shape to some piece‐wise ellipses. A cost function is defined to detect the cell contours incorporating both the elliptical shape and local image intensities. The procedure is fast, reliable and well suited for routine interactive applications.

Nuclear diffuseness as a measure of texture: Definition and application to the computer‐assisted diagnosis of parathyroid adenoma and carcinoma

A measure of texture, the nuclear diffuseness, was formulated for use in biological classification, and specifically to characterize quantitatively chromatin texture. Nuclear diffuseness corresponds to the amount of local intensity variation in the digitized image of a nuclear profile. As a setting in which to test the efficacy of nuclear diffuseness as a diagnostic tool, the identification of parathyroid adenoma and carcinoma was considered. Digitized images of sections of parathyroid chief cell nuclei were obtained from 16 biopsies, and the nuclear diffuseness, as well as other morphometric descriptors, were computed. With just the average nuclear diffuseness and average nuclear profile area, jackknife (leave‐one‐out) classification using an artificial neural network was able to diagnose correctly and unambiguously the condition (normal, parathyroid adenoma, or parathyroid carcinoma) in 15 of 16 cases. In one case, the neural network assigned a higher weight to the correct diagnosis, but was unable to distinguish between normal and adenoma conclusively.

Minimum entropy restoration of star field images

In this correspondence, we present an algorithm for restoration of star field images by incorporating both the minimum mean square error and the maximum varimax criteria. It is assumed that the point spread function of the distortion system can be well approximated by a Gaussian function. Simulated annealing (SA) is used to implement the optimization procedure. Simulation results for both Gaussian and square point spread functions with heavy additive independent white Gaussian noise are provided. Visual evaluation of the results indicate that the proposed algorithm performs better than the noncausal Wiener filtering method.

An iterative algorithm for cell segmentation using short-time Fourier transform.

In this paper, an iterative cell image segmentation algorithm using short‐time Fourier transform magnitude vectors as class features is presented. The cluster centroids of the magnitude vectors are obtained by the K‐means clustering method and used as representative class features. The initial image segmentation classifies only those image pixels whose surrounding closely matches a class centroid. The subsequent procedure iteratively classifies the remaining image pixels by combining their spatial distance from the regions already segmented and the similarities between their corresponding magnitude vectors and the cluster centroids. Experimental results of the proposed algorithm for segmenting real cell images are provided.

The use of local entropy measures in edge detection for cytological image analysis

Accurate edge detection is a fundamental problem in the areas of image processing and pattern recognition/classification. The lack of effective edge detection methods has slowed the application of image processing to many areas, in particular diagnostic cytology, and is a major factor in the lack of acceptance of image processing in service orientated pathology. In this paper, we present a two‐step procedure which detects edges. Since most images are corrupted by noise and often contain artefacts, the first step is to clean up the image. Our approach is to use a median filter to reduce noise and background artefacts. The second operation is to locate image pixels which are ‘information rich’ by using local statistics. This step locates the regions of the image most likely to contain edges. The application of a threshold can then pin‐point those pixels forming the edge of structures of interest. The procedure has been tested on routine cytologic specimens.

A focusing algorithm for high magnification cell imaging

An algorithm to produce a uniformly focused image in digital acquisition of high magnification light microscopy images is presented. In very high magnification microscopic imaging the specimen surface cannot be considered ideally flat so that capturing a single image frame is usually not sufficient to capture an image that is focused everywhere. An image formation model for light microscopic images is presented, and based on this model an algorithm to construct a uniformly focused image is presented. The algorithm requires that multiple frames of the image at different focal planes be processed to combine their information to obtain an estimated of the desired image which is more completely focused than any of the individual frames. Experimental results show that the proposed algorithm is very effective in approximating the desired image in high magnification microscopic imaging and highly robust comparing to the gradient method.

3D Arterial Trace Reconstruction From Biplane Multi-Valued Projections

An automatic algorithm for reconstructing arterial center lines in three dimensional (3D) space from two orthogonal angiographic views is presented. As a result of representing projected center lines by, cubic spline polynomials, corresponding points in both views are automatically determined. A previous paperl showed automatic positional reconstruction to be possible when the projected center line can be expressed as a single-valued function. This algorithm generalizes the method to include cases where the center lines are described by multi-valued functions. Three dimensional curves, representing arterial center lines, were sampled and projected onto two orthogonal planes to simulate the projected vessel center line in each view. Gaussian noise of different magnitudes was added to the projected coordinates in both views to simulate vessel center line estimation errors. Stenosed segments were simulated by deleting sections of the projected center lines. Positional reconstruction accuracy for various mean centering errors (MCE) and stenosis lengths are presented.

An Approach to Integrating Health Disparities within Undergraduate Biomedical Engineering Education

Health disparities are preventable differences in the incidence, prevalence and burden of disease among communities targeted by gender, geographic location, ethnicity and/or socio-economic status. While biomedical research has identified partial origin(s) of divergent burden and impact of disease, the innovation needed to eradicate health disparities in the United States requires unique engagement from biomedical engineers. Increasing awareness of the prevalence and consequences of health disparities is particularly attractive to today’s undergraduates, who have undauntedly challenged paradigms believed to foster inequality. Here, the Department of Biomedical Engineering at The City College of New York (CCNY) has leveraged its historical mission of access-and-excellence to integrate the study of health disparities into undergraduate BME curricula. This article describes our novel approach in a multiyear study that: (i) Integrated health disparities modules at all levels of the required undergraduate BME curriculum; (ii) Developed opportunities to include impacts of health disparities into undergraduate BME research projects and mentored High School summer STEM training; and (iii) Established health disparities-based challenges as BME capstone design and/or independent entrepreneurship projects. Results illustrate the rising awareness of health disparities among the youngest BMEs-to-be, as well as abundant undergraduate desire to integrate health disparities within BME education and training.