Low-quality medical image contours in nonsubsampled contourlet domain

Abstract

Active contour model is one of the popular methods for computer vision. In medicine, a contour presents the shape and the size of parts of the body. It is useful for clinical diagnosis of medical specialists in the treatment of patients. However, most medical images have many noise types in the real world. This is a big challenge for the contour determination. In this paper, we propose a new method to find contours in medical images which have noise by applying active contour model in nonsubsampled contourlet domain. Our algorithm improves the ability for smoothing before reducing energy between boundaries detected in nonsubsampled contourlet domain. Compared with other recent methods, the proposed method is better.