Pre-Processing Method for Extraction of Pectoral Muscle and Removal of Artefacts in Mammogram

Mammography is an effective imaging modality used by radiologists for detection of breast cancer and effective suppression of pectoral muscle as well as removal of noise in form of artefacts aids accurate detection. In existing methods, either over-segmentation or under-segmentation, have been reported resulting in less accurate results. In this work, an algorithm was developed that effectively removed artefacts and pectoral muscle from mammograms, it uses a region description method, which uses splitting and merging technique to extract pectoral muscle and remove unwanted noise from the original mammograms. Digital mammograms from MIAS database were used to test the algorithm. The algorithm achieved 92.24% good results,4.04% acceptable results,3.10% unaccepted results, and 0.62% of the mammograms were discarded. Experimental results obtained indicate that the breast regions extracted accurately. Keywords: Mammograms, Pre-processing, Region of Interest (ROI).