Date of Award


Document Type


Degree Name

Master of Science


Department of Electrical and Computer Engineering

First Advisor

Steven K. Rogers, PhD


An automated method for detecting microcalcification clusters is presented. The algorithm begins with a digitized mammogram and outputs the center coordinates of regions of interest (ROIs) that contain microcalcification clusters. The method presented uses a 12-tap Least Asymmetric Daubechies (LAD12) wavelet in a tree structured filter bank to increase the signal to noise level of microcalcifications. The signal to noise level gain achieved by the filtering allows subsequent thresholding to eliminate on average 90% of the image from further consideration without eliminating actual microcalcifications 95% of the time. A novel approach to isolating individual calcifications from background tissue through non-stationary noise reduction, low/hi thresholding, and morphological filtering is demonstrated; this technique reduces the number of false detections by an average of 5 per image. Several features are extracted from each potential calcification, including two newly proposed correlation features, to distinguish actual microcalcifications from correlated background tissue. Altogether, the method successfully detected 44 of 53 microcalcification clusters (83%) with an average of 2.3 false positive clusters per image. A cluster is considered detected if it contains 3 or more microcalcifications within a 6.4 mm by 6.4 mm area. Although the emphasis is placed on detecting microcalcification clusters for further examination by a radiologist with no attempt made to diagnose the cluster as malignant or benign, the method successfully detected 13 of 14 (93%) malignant cases.

AFIT Designator


DTIC Accession Number