. Detecting Region-of-Interest (ROI) in Digital Mammogram by using
Jucheng Yang, Dongsun Park · 2004
Microcalcificarions are early sign of breasr cancer appear as isolared bright spors in mammogram images, which are dificult ro detecr due fo their small size. In this paper, morphological bandpass filter(MBF) is introduced to derecf microcalcificarions, which is implemented by opening rhe original image WO times with two different strucrure elements respectively, and subrracting one opened image using another one rhar can decompose the image info interest derails domain image where microcalcifications rend ro appear. Series of MBFs are tuned for rhe detection task, and binary image conrained microcalcifications region-of-interest (ROI) will be obrained. Experimental results show that the proposed method wirh these bandpass filters can recognize ROI wirh a rrue posirive of 93.07% and a false positive of 4.34%. Comparing to the well-known discrere waveler rransfom (OW) method, this method is more accurate in positions and sizes of microcalc@carions.