R. Chandrasekhar

A simple method for automatically locating the nipple on mammograms

This paper outlines a simple, fast, and accurate method for automatically locating the nipple on digitized mammograms that have been segmented to reveal the skin-air interface. If the average gradient of the intensity is computed in the direction normal to the interface and directed inside the breast, it is found that there is a sudden and distinct change in this parameter close to the nipple. A nipple in profile is located between two successive maxima of this parameter; otherwise, it is near the global maximum. Specifically, the nipple is located midway between a successive maximum and minimum of the derivative of the average intensity gradient; these being local turning points for a nipple in profile and global otherwise. The method has been tested on 24 images, including both oblique and cranio-caudal views, from two digital mammogram databases. For 23 of the images (96%), the rms error was less than 1 mm at image resolutions of 400 /spl mu/m and 420 /spl mu/m per pixel. Because of its simplicity, and because it is based both on the observed behavior of mammographic tissue intensities and on geometry, this method has the potential to become a generic method for locating the nipple on mammograms.

Automatic pectoral muscle segmentation on mammograms by straight line estimation and cliff detection

Mammograms, which are X-ray images of the female breast, are used widely by radiologists to screen for breast cancer. The first stage of any computerized analysis of the digitised mammogram is to divide the image into anatomically distinct regions. The pectoral muscle is one of these regions and it appears on mediolateral oblique views of mammograms. In this paper, the rationale and algorithms for fully automatic, two-part segmentation of the pectoral muscle are presented. The algorithm consists of (a) estimation of the muscle edge by a straight line; and (b) refinement of the detected edge by surface smoothing and edge detection in a restricted neighbourhood derived from the first estimate.

Gross segmentation of mammograms using a polynomial model

The breast and background on a mammogram form complementary, connected sets. Generally, the intensities comprising the background are spatially continuous, low in value and lie within a closed interval. The background may therefore be approximated by a polynomial in x and y on the basis of the Weierstrass approximation theorem. The authors include the whole background and a small portion of the breast in the region being modelled. The modelled background is subtracted from the original image, the resulting image thresholded, and the largest low intensity region taken to be the background. Connected regions are identified, labelled and merged. The background is floodfilled, and inclusions removed from the object, to yield a breast-background binary image. The method has been tested on 58 mammograms of two views from two digital mammogram databases. With one exception, it performs well and yields a skin-air interface with sufficient fidelity to preserve a nipple in profile.

Spatially based application of the minimum cross-entropy thresholding algorithm to segment the pectoral muscle in mammograms

A threshold-based algorithm is presented for the extraction of the pectoral muscle edge in mediolateral oblique view mammograms. The minimum cross-entropy thresholding algorithm is applied to local areas around the pectoral muscle to determine a series of thresholds as a function of area size. Using a model image it is shown that art inflection point in this function corresponds to a threshold that will separate the pectoral muscle from the rest of the breast. Post processing is performed on mammograms to eliminate false positive points of inflection and a straight line is fitted to the detected pectoral boundary in order to smooth jaggedness caused by the non-uniform intensity of the pectoral muscle edge.

Automatic assessment of mammographic positioning on the mediolateral oblique view

Mammograms are X-ray images of the compressed breast and are widely used for early detection of breast cancer. A mammogram must be of sufficient quality for the radiologist to detect lesions or other abnormalities with high sensitivity and specificity. Some algorithms are presented for the automatic assessment of the quality of positioning on mediolateral oblique (MLO) view mammograms. Anatomic features, including the breast border, nipple location and pectoral margin, were first extracted from each image. Then several quality criteria, including breast tissue exclusion, nipple in profile, inclusion of inframammary fold, and positioning of the pectoral muscle, were used to assess the adequacy of breast positioning. The assessment method was tested on 322 digitized mammograms in the MIAS database.

DSP in mammography

Breast cancer is the most frequently occurring cancer in females with no cure at present. Early detection offers the best chance of survival and mammography is used to screen the asymptomatic female population above fifty years of age. Computerized analysis of mammograms can assist radiologists to detect lesions or abnormalities. However, the entire digitized mammogram must first be segmented and analyzed prior to lesion detection. This paper summarizes work done at the Australian Research Centre for Medical Engineering (ARCME) to systematically and hierarchically segment mammograms as a precursor to lesion detection. The breast is first segmented from the non-breast background by polynomial modelling and subtraction of the latter region. The nipple, which is the only anatomical landmark, is then located using a sensitive feature set to search the breast border. The pectoral muscle is then identified by an adaptive edge detection/surface fitting algorithm. Finally, a systematic methodology is proposed for lesion search.

Semi-Automatic Tracking of the Diaphragm Contour in X-Ray Image Sequences: Preliminary Results

A semi-automatic method is described for tracking the contour of the diaphragm in lateral X-ray image sequences. It is part of a method being developed to quantify diaphragm function non-invasively. Initialization is achieved by interactive segmentation of the diaphragm contour and identification of landmarks. The method relies on modelling the contour using an augmented active contour model incorporating anatomical constraints. The contour is then tracked iteratively across successive frames using reduced-search space dynamic programming. The method has been tested on several image sequences and preliminary results are promising with an 85.3 % tracking accuracy for one such sequence.

Unconventional edge detector: preliminary theoretical investigation

We have previously described a range-based neighbourhood operator and an experimentally discovered unconventional edge detector based on it. The latter relies on data fitting in a pixel neighbourhood, and has a wide dynamic range. A preliminary theoretical investigation of its basis, and some of its properties, are presented in this paper. It is revealed that the edge sensitive feature is the range of pixel values in successive pixel annuli around a central pixel. It is also shown that the edge strength at any centre pixel may be approximated, in the first instance, by the sum of the logarithms of these annular range values. The derived approximation is illustrated with a synthetic image and a mammogram.

Analysis of digitized lateral fluoroscopy images to quantify the volume displaced by the diaphragm

Methods have previously been been developed for accurately measuring the volume displaced by motion of the diaphragm (/spl Delta/Vdi) and the contribution of the diaphragm to inspired volume. These methods require accurate measurements of (i) the surface area swept by the diaphragm during inspiration, (ii) the area within this that is occupied by the vertebral column and associated tissues, and (iii) the diameter of the lower rib cage. Existing methods of measurement are labor intensive and slow. We present a more accurate and efficient system for the acquisition, distortion correction, interactive segmentation and functional analysis of fluoroscopic images. This new non-invasive technique will allow earlier and more accurate detection of abnormal diaphragm function and have a direct clinical application.

A graphical interface for viewing mammograms interactively

Presents the design principles for a graphical user interface for viewing mammograms interactively, in the context of self-paced, computer-aided instruction. Patient data, whole image and full resolution views, feature highlighting, image processing and radiology and pathology reports, are all integrated into a self-contained package with intuitively understood graphical icons to permit rapid learning and comfortable use. The image database is searchable by patient, view, date, similarity of lesion appearance, and pathology so that different slices of the same data may be reviewed to consolidate knowledge and test understanding.