Mammogram Classification Based on Morphological Component Analysis (MCA) and Curvelet Decomposition

Author(s): Syed J.S. Gardezi, Ibrahima Faye.

Journal Name: Neuroscience and Biomedical Engineering (Discontinued)

Volume 3 , Issue 1 , 2015

Graphical Abstract:


Abstract:

Detection of breast cancer at early stages helps reducing the mortality rates in women. Mammography has proven to be very useful tool in the diagnosis of breast cancer, yet there are complications to separate diverse morphological features in mammographic images. This study investigates the potential use of morphological component analysis (MCA) for classifying normal and abnormal tissues in mammograms. Two different dictionaries i.e. Local discrete cosine transform (LDCT) and Curvelet transform via wrapping (CURVwrap) were used with varying iterations to obtain a morphological decomposition into two parts i.e. a piecewise smooth part and a texture part. The piecewise smooth parts for each iteration were kept and features were extracted using curvelet coefficients. Using the Simple logistic classifier with 10-folds cross validation, an accuracy of 86.50% with AUC value of 0.89 was achieved for 300 iterations.

Keywords: Breast cancer, classification component, curvelet transform, morphological component analysis (MCA).

Rights & PermissionsPrintExport Cite as


Article Details

VOLUME: 3
ISSUE: 1
Year: 2015
Page: [27 - 33]
Pages: 7
DOI: 10.2174/2213385203666150619175230

Article Metrics

PDF: 15