Masses automatic classification in dense breast images by wavelets transform (ref. 31,37)

Development of an automatic classifier scheme based on the wavelet transform for masses identified in digitized mammographic images. This classifier is part of a mammography CAD scheme. A minimum distances classifier was also used in order to distribute the images in their respective classes. The results have shown that the classifier is able to differentiate suspect and non-suspect nodules with more than 90% of right rate.