The potential of mathematical morphology for contour extraction from ultrasound images

This paper presents two applications for contour extraction from ultrasonic images. The first is the extraction of the endocardial boundaries from a sequence of echocardiographic images, and the second is the extraction of the contours of follicles in the ovary from an ultrasound obstetrical image. Mathematical morphology shows its potential for filtering and segmenting the images. This is because mathematical morphology is very efficient in dealing with geometrical features such as size, shape, contrast, and connectivity that can be considered as segmentation-oriented features. Moreover, morphological filters by reconstruction have a size-oriented simplification effect on the signal but preserve the contour information. Although linear filters are more pervasive, they are by no means ideal for ultrasound B-mode images