Extraction of microcalcifications from digital mammograms using mathematical morphology

It is often difficult to detect and interpret subtle abnormalities on mammograms. Interpretation by radiologists may be aided by digitisation of the films followed by application of computer based image analysis techniques. Microcalcifications represent the commonest if not the only radiographic proof of early breast cancer. Between 20 and 25% of breast carcinomas detected radiographically show clustered microcalcifications and any increase in the detection of microcalcifications is likely to improve the detection of breast cancer. The present study investigates the use of mathematical morphology algorithms for the identification and segmentation of microcalcifications and extraction of numerical information about them