Joubin Hatamzadeh

Segmentation of objects within images is one of the most important problems in computer vision and image processing. As a high-level technique for segmentation, active contours or snakes are used extensively for boundary identification of objects in images. When used as the image force for active contours, the gradient has the disadvantage of having a restricted capture range. Two solutions for improving the capture range, gradient vector flow (GVF) and pressure force, are compared. Although GVF provides a good capture range, it sometimes wrongly identifies boundaries. As an alternative, pressure forces have shown promising results for histological and MRM middle-ear images. The use of open contours is also demonstrated, in addition to the usual closed contours.