Application of weighted-majority minimum-range filters in the detection and sizing of tumors in mammograms

摘要

Abstract: In image processing the solution is often unique to theproblem. To be more specific, the importance of thefilter window and sampling pattern chosen to filter,pass, or enhance a specific shape is very specific tothe problem at hand. We detect suspect tumors inmammograms using a weighted majority minimum rangefilter and different sampling patterns and windows as ademonstration of this fact. Several methods have beendeveloped to automate the process of detecting tumorsin mammograms. We show that traditional windowing orsampling methods may be replaced by a hexagonal methodthat more accurately reflects the geometry of theproblem and could improve the techniques already inexistence. Several theorems involving a hexagonalfilter window are presented, followed by the results ofour application to mammograms. !13