Computer Aided Diagnosis: Automated detection
and enhancement of microcalcifications
in digitized mammograms using wavelet decomposition
and local gray thresholding
Détection automatique des microcalcifications
dans les mammogrames digitales utilisant
N. Hamdi1, K. Auhmani1,2*, M. M. Hassani1
1 LEI, Département de Physique, Faculté des sciences Semlalia, Université Cadi Ayyad, Marrakech 40000 Maroc,
2 Ecole Nationale des sciences appliquées, BP 63 Safi 46000, Université Cadi Ayyad, Maroc,
* Corresponding author. E-mail: firstname.lastname@example.org
Received: 30 April 2008; revised version accepted: 18 September 2008
Microcalcifications are suspected to be among the first signs of breast cancer. When designing a computer system for analysis of mammograms, it is necessary to find methods suitable for locating microcalcifications. We have developed a method for the detection of microcalcification clusters in digitized mammogram images. This is a multi step process. One is an adaptive mammogram enhancement algorithm using homomorphic filtering and wavelet enhancement of objects occurring at scales characteristic of microcalcifications. The microcalcifications correspond to high frequency component of the image spectrum. Detection of microcalcifications is achieved by decomposing the mammogram into different frequency subbands, suppressing the low frequency subbands, and finally reconstructing the mammogram from the subbands containing only high frequency. The other step is local thresholding of the enhanced image. We have tested the proposed system on the MIAS database. Data analysis shows that the proposed method can effectively detect the occurrence of microcalcifications.
Keywords: Mammography; Wavelet transforms; Enhancement; Homomorphic filter; Opening; Local thresholding.