A method for detecting microcalcifications in Digital Mammograms

Abstract

Microcalcification clusters are often an important indicator for the detection of malignancy in mammogratas. In many cases, microcalcifications are the only indication of a malignancy. However, the detection of microcalcifications can be a difficult process. They are small and can be embedded in dense tissue. This paper presents a method for automatically detecting microcalcifications. We utilize a high-boost fUter to suppress background clutter enabling segmentation even in very dense breast tissue. We then use a threshholding and region growing technique to extract candidate microcalcifications. Likely microcalcifications are then identified by a linear classifier. We apply this method to images selected from the LLNL/ UCSF Digital Mammogram Library, and produce a receiver operating characteristic (ROC) curves to detail the trade-off between probability of detection and false alarms. Finally, we exam the ability to properly select a threshold to achieve a desired probability of detection based upon a training set. This is a US government work. There are no restrictions on its use.KEY WORDS: digital mammography, microcalcifications, high-boost filtering, detection. MICROCALCIFICATIONS ARE ah important early indicator of malignancy in mammography. Clusters of microcalcifications are present in between 30% and 50% of breast cancers, and such clusters are the only visible sign in approximately 36% of these cases. ~ Unfortunately, microcalcifications are often difficult to locate. They ate generally smaller than 1.0 mm in diameter, and they are usually in the 0.1 to 0.3 mm range. As few as five small microcalcifications can be an indication of malignancy. Furthermore, microcalcification clusters can occur in dense breast tissue where they ate not easily discernible. Asa result, it is often necessary for a radiologist to carefully scan a

New Dynamic Self-Organizing Feature Maps for the classification of extracted feature vectors of characters

“…The brief idea of the algorithm is that every image's pixel is set to black if its brightness is lower than the average brightness of surrounding pixels in the window of the specified size (specifies window size around processing pixel, which determines number of neighbour pixels to use for calculating their average brightness), otherwise it is set to white. Suppose the illumination over a handwritten file is non-uniform, for example in the case of some scanned pages of book or camera captured document files, global binarization methods tend to produce marginal noise along the page borders [4]. Another class of files to be processed is historical document files in which the intensities of image can be changed significantly within a file.…”

Computational adjustment technique for digital mammographic images based on the digitizer characteristic curve

“…15 The purpose here is to evaluate the effect of the digitization procedure regarding these technologies on the sensitivity and specificity of a procedure designed to detect clustered microcalcifications in digitized mammograms. The segmentation procedure is based on the combination of algorithms developed previously by Nishikawa et al 16 and Wallet et al, 17 with the purpose of automatically detecting microcalcifications in digitized breast images, tested in some works, with wellknown results. The main reason for joining techniques was that one missed microcalcifications that were however detected by the other and vice versa for many mammography images and regions of interest ͑ROIs͒.…”

Contrast enhancement in dense breast images using the modulation transfer function

“…However, identifying clustered microcalcifications in visual inspection of mammograms is a hard task, particularly for images of dense breasts in which contrast between microcalcifications and tissues is poor. Dense breast images are also a challenge for computeraided diagnosis ͑CAD͒ schemes, 5,6 again due to the poor contrast. Preprocessing techniques are therefore required to enhance contrast and make the images suitable to the conventional CAD scheme processing.…”

“…Exceptions are represented by three pieces of work, as follows. Wallet et al 5 proposed a high-boost filter to eliminate the mammogram's background and thus prepare the image for segmentation of the microcalcifications. They state that using such a filter is better than employing techniques using thresholding, mainly in dense breasts images.…”