A discrete dynamic contour model

Abstract

A discrete dynamic model for defining contours in 2-D images is developed. The structure of this model is a set of connected vertices. With a minimum of interaction, an initial contour model can be defined, which is then automatically modified by an energy minimizing process. The internal energy of the model depends on local contour curvature, while the external energy is derived from image features. Solutions are presented to avoid undesirable deformation effects, like shrinking and vertex clustering, which are common in existing active contour models. The deformation process stops when a local minimum of the energy function is reached. The final shape of the model is a reproducible approximation of the desired contour. Results of applying the method to computer-generated images, as well as clinical images, are presented.

Group Average Difference: A Termination Criterion for Active Contour

“…Taking unity magnitude for normal force makes the contour move in a more consistent way which is dominated by the external forces relative to the shape of the contour. To further avoid vertex dispersion and clustering problem, contour resampling operation is carried out (but not for evolutions terminated based on pixel movement) for each time step of the evolving contour, as suggested by [14].…”

“…Contour resampling is the addition and removal of contour vertices in subsequent iterations, as suggested in [14]. The contour resampling step is necessary for avoiding vertex clustering and vertex dispersion.…”

Optimal Weights for Convex Functionals in Medical Image Segmentation

“…The result is a convex function in w under a linear constraint, since the first term in (2) has been shown to be a convex quadratic [9] and the second term linear. A convex function under a linear constraint can be solved via convex optimization, and thus the optimal w(I j ) is guaranteed.…”

“…For this and other reasons, the automatic segmentation of medical images remains a daunting task. Many segmentation approaches rely on the minimization of objective functions, including several landmark papers: from the seminal paper of Snakes for 2D segmentation [1] and other explicit models [2] to implicit models [3,4], graph approaches [5,6], and variants thereof.…”

Physically And Statistically Based Deformable Models For Medical Image Analysis

“…One approach for dealing with such objects is the use of deformable models. Deformable models [59] such as snakes [2] and its variants [16,[42][43][44][45], have attracted considerable attention and are widely used for segmenting non-rigid objects in 2D and 3D (volume) images. However there are several well-known problems associated with snakes.…”