Tissue density classification in mammographic images using local features

In breast cancer cases, it is known that the ratio of correct diagnosis is affected by the breast tissue density. For this reason, automatic tissue density classification is an important process in diagnosis. In this work a method for classification of breast tissue density from mammographic images is proposed. The objective of the method is to determine which class, namely fatty, fatty-glandular and dense-glandular, the breast tissue belongs to. For this purpose, SIFT algorithm is used as the local feature extraction method, and LVQ algorithm is used for supervised classification. Test results on the MIAS dataset demonstrate that the code vectors corresponding to bag of SIFT features of each class can successfully model the breast tissue and the classification accuracy over 90% is achieved by LVQ.

[1]  Ian W. Ricketts,et al.  The Mammographic Image Analysis Society digital mammogram database , 1994 .

[2]  G LoweDavid,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004 .

[3]  Styliani Petroudi,et al.  Combination of different texture features for mammographic breast density classification , 2012, 2012 IEEE 12th International Conference on Bioinformatics & Bioengineering (BIBE).

[4]  Arnau Oliver,et al.  Automatic classification of breast density , 2005, IEEE International Conference on Image Processing 2005.

[5]  Wei Wang,et al.  A review of breast tissue classification in mammograms , 2011, RACS.

[6]  Ian H. Witten,et al.  The WEKA data mining software: an update , 2009, SKDD.

[7]  M. V. Velzen,et al.  Self-organizing maps , 2007 .

[8]  Ying Zhang,et al.  Bag-of-Features Based Classification of Breast Parenchymal Tissue in the Mammogram via Jointly Selecting and Weighting Visual Words , 2011, 2011 Sixth International Conference on Image and Graphics.

[9]  Arnau Oliver,et al.  Modeling and Classifying Breast Tissue Density in Mammograms , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).