This work proposes a method aimed at enhancing the contrast in dense breast images in mammography. It includes a new preprocessing technique, which uses information on the modulation transfer function (MTF) of the mammographic system in the whole radiation field. The method is applied to improve the efficiency of a computer-aided diagnosis (CAD) scheme. Seventy-five regions of interest (ROIs) from dense mammograms were acquired in two pieces of equipment (a CGR Senographe 500t and a Philips Mammodiagnost) and were digitized in a Lumiscan 50 laser scanner. A computational procedure determines the effective focal spot size in each region of interest from the measured focal spot in the center for a given mammographic equipment. Using computational simulation the MTF is then calculated for each field region. A procedure that enlarges the high-frequency portion of this function is applied and a convolution between the resulting new function and the original image is performed. Both original and enhanced images were submitted to a processing procedure for detecting clustered microcalcifications in order to compare the performance for dense breast images. ROIs were divided into four groups, two for each piece of equipment-one with clustered microcalcifications and another without microcalcifications. Our results show that in about 10% of the enhanced images more signals were detected when compared to the results for the original dense breast images. This is important because the usual processing techniques used in CAD schemes present poor results when applied to dense breast images. Since the MTF method is a well-recognized tool in the evaluation of radiographic systems, this new technique could be used to associate quality assurance procedures with the processing schemes employed in CAD for mammography.

The performance of a computer-aided dianosis (CAD) scheme is closely dependent on the database used for its development and tests. The scheme sensitivity can be reduced by 15% to 25%, with only 20% of changes in the database cases. Previously, we have developed a processing scheme in order to detect clustered mlcrocalcifications in digital mammograms, and we have tested such a procedure with two different databases. Further evaluations in developing a CAD scheme for mammography have indicated the need for more extensive investigation on the effects resulting from different characteristics of the images bank used for tests. Therefore, this work reports some results regarding such an investigation, with a further discussion over characteristics that can affect the performance of a CAD scheme. Copyright © 2001 by W.B. Saunders CompanyT HE PERFORMANCE of a computer-aided diagnosis (CAD) scheme is strongly dependent on the database used for testing the procedures. I In a previous work? we developed a computerized procedure as part of a mammography CAD scheme intended to detect clustered microcalcifications in digitized mammograms. The first tests with this scheme indicated an efficacy rate of 94% for a particular set of images. With the goal to study the effect of the characteristics of different sets of images on the performance of such a procedure, we have checked the results yielded from five different sets of mammograms. METHODSThe following sets of images were used for the investigation: (I) a set of actual mammograms, obtained from the archives of Hospital das Clfnicas . same set digitized by a LUl\lISCAN (Lumisys, Sunnyvale, CA) laser scanner, with 0.15 mm and 8 bits; (3) another set of mammograms obtained from that same hospital in 1995, and digitized in a UMAX UCI260-Pro, with the same resolutions as the first group; (4) a set of mammograms obtained from recent examinations performed at the Santa Casa hospital at S. Carlos (Brazil), only corresponding to dense breast cases and digitized by the LUl\lISCAN scanner, with the same resolutions as the second group above; and (5) a previously used set obtained off the internet from the National Expert and Training Center for Breast Cancer Screening (the University of Nijmegen, the Netherlands), also with the same resolutions as the second and fourth groups.The steps of the processing scheme for clusters detection were as follows: identification of regions of interest (ROIs); ROI segmentation; area-point transformation.' which converts each detected signal in a unique pixel; and microcalcification grouping, where clusters are identified and marked in the final image. Initially, the procedure was applied to all of the image sets with no changes in the processing parameters, in order to investigate the effect of the acquisition and digitization processes. Then, some parameters were changed according to the database characteristics under test in order to determine the best result possible for that set of images. Figure I illustrates the resultant images after some step...

No abstract

No abstract

This article describes a segmentation technique that uses the watershed transformation to provide mass detection in dense breast digitized images. The technique consists of four steps: preprocessing, which isolates the breast from the image background; histogram equalization to enhance the contrast; watershed transformation, which calculates the gradient for the whole image; and a topography technique that organizes the segmented parts. We processed 109 regions of interest extracted from mammograms. Preliminary results showed 20% false positive and 85.4% true positive detection regarding the equalized images containing masses. Good results were obtained with this technique. However, additional techniques should be developed to eliminate noise in some images. This technique could be an important tool in mammography.

Dense breasts, that usually are characteristic of women less than 40 years old, difficult many times early detection of breast cancer. In this work we present the application of some image processing techniques intended to enhance the contrast in dense breast images, regarding the detection of clustered microcalcifications. The procedure was, firstly, determining in the literature the main techniques used for mammographic images contrast enhancement. The results indicate that, in general: (1) as expected, the overall performance of the CAD scheme for clusters detection decreased when applied exclusively to dense breast images, compared to the application to a set of images without this characteristic; (2) most of the techniques for contrast enhancement used successfully in generic mammography images databases are not able to enhance stmctures of interest in databases formed only by dense breasts images, due to the very poor contrast between microcalcifications, for example, and other tissues. These features should stress, therefore, the need of developing a methodology specifically for this type of images in order to provide better conditions to the detection of breast suspicious stmctures in these group of women.

No abstract

No abstract