Automated extraction of microcalcifications BI-RADS numbers in mammograms

Author

Quadrades, Salvador ; Sacristán, Amparo

Author_Institution

Telecomunications & Archit. of Comput. Dept., Mataro Sch. of Eng., Barcelona, Spain

Volume

2

fYear

2001

Firstpage

289

Abstract

A detection scheme for microcalcifications and automated BI-RADS numbers is presented. Pixels suspected to be microcalcifications are detected using morphological image analysis with mathematical morphology. The study is divided into three separate parts. First, the regions of interest (ROI) are located, second the ROI are segmented, and finally the microcalcification BI-RADS numbers (calcification distribution, calcification number, and calcification description) are extracted. Results show a good microcalcification detection (100% of it), segmentation (about 99%) and the innovation of automatization of BI-RADS numbers is in concordance with radiologist values

Keywords

cancer; feature extraction; image segmentation; mammography; mathematical morphology; medical image processing; object detection; BI-RADS numbers; ROI segmentation; automated extraction; breast cancer; calcification description; calcification distribution; calcification number; feature extraction; mammograms; mathematical morphology; microcalcifications; morphological image analysis; pixel detection; regions of interest; Breast cancer; Computer architecture; Gaussian distribution; Image analysis; Image databases; Image reconstruction; Image segmentation; Morphology; Pixel; Telecommunication computing;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 2001. Proceedings. 2001 International Conference on

Conference_Location

Thessaloniki

Print_ISBN

0-7803-6725-1

Type

conf

DOI

10.1109/ICIP.2001.958484

Filename

958484