Automatic mass segmentation method in mammograms based on improved VFC snake model - Emerging Trends in Image Processing, Computer Vision, and Pattern Recognition - page 165

2.2.3 Morphological filter
Then we amend the enhanced image using morphological filter. The preprocessed image is
shown in Figure 2 .
FIGURE 2 Image(from MIAS) before and after enhancement and preprocessed results.
2.3 ROI Extraction and Location
2.3.1 Edge extraction
The preprocessed image is composed of pectoral muscle and mass region. Before removing
the pectoral muscle, the edge detection operator is used to extract the edge first.
2.3.2 Hough transform detection
From Figure 3 , the edge of pectoral muscle appears as a triangle while mass edge usually ap-
pears as an ellipse or circle; thus, we can obtain the approximate edge of mass by performing
the linear Hough transform (LHT) [ 18 ] and circular Hough transform (CHT) on the extracted
edge image, which, respectively, are defined as formulas (6) and ( 7 ) . Here, Ω pectoral is the pec-
toral muscle detected by LHT on edge image, Ω mass is the initially segmented mass obtained
by CHT on the region Ω edge-pectoral , where Ω pectoral is removed. ( a i , b i ) is the center position of the
circle and R is the radius. The location results are shown in Figure 3 :
(6)
(7)