Automatic detection of abnormal tissue in mammography

Abstract

A novel method for accurate detection of regions of interest (ROIs) that contain circumscribed lesions in mammograms is presented. The mammograms are segmented using a statistical threshold and a number of candidate regions are extracted. Then a set of qualification criteria is employed to filter these regions retaining the most suspicious for which a Radial-Basis Function Neural Network makes the final decision marking them as ROIs that contain abnormal tissue. The proposed method detects the exact location of the circumscribed lesions with accuracy of 90.9%, and a very low number of false positive regions per image (2.1 ROIs per image) in the MIAS database.

An Approach with Support Vector Machine using Variable Features Selection on Breast Cancer Prognosis

“…The best specificity is achieved when attribute A [2,3,4,5,6,7,8,9] are selected for training and testing with 0.99163 and 1.00 respectively. The lowest error rate and the best AUC are obtained with A [2,3,4,5,6,7]. The accuracy is the proportion of the total number of predictions that were correct.…”

“…As shown in table 3 the classifier gives the best sensitivity with 0.9886 0.9889 with attribute A [2,3,4,5,6,7] are selected for training and testing the machine for each 10 and 5 cross validation respectively. The best specificity is achieved when attribute A [2,3,4,5,6,7,8,9] are selected for training and testing with 0.99163 and 1.00 respectively. The lowest error rate and the best AUC are obtained with A [2,3,4,5,6,7].…”

“…The accuracy is the proportion of the total number of predictions that were correct. The best accuracy is evaluated when we considered the attribute A [2,3,4,5,6,7] …”

“…The etiologist of breast cancer remain unclear and no single dominant cause has emerged. [2] [3] Prevention is still a mystery and the only way to help patients survive is by early detection. If the cancerous cells are detected before spreading to other organs, the survival rate for patients is more than 97%.…”

Toward a standard reference database for computer-aided mammography

“…For instance, Zwiggelaar et al used synthetic data and 15 mammographies of the Mammographic Image Analysis Society digital mammogram database (MIAS, http://peipa.essex.ac.uk/ipa/pix/mias/) to detect linear structures and classify them into the anatomical types: vessels, ducts, and spicules [4]. Christoyianni et al detected the exact location of circumscribed masses in 22 images of MIAS database [5]. Strickland & Hahn employed wavelet decomposition in 40 mammographies of Nijmegen database to detect micro-calcifications [6].…”