Limits...
New region-scalable discriminant and fitting energy functional for driving geometric active contours in medical image segmentation.

Wang X, Niu Y, Tan L, Zhang SX - Comput Math Methods Med (2014)

Bottom Line: The region-scalable discriminant and fitting energy functional is defined to capture the image intensity characteristics in local and global regions for driving the evolution of active contour.The discriminant term in the model aims at separating background and foreground in scalable regions while the fitting term tends to fit the intensity in these regions.The new model utilizes intensity information in the local and global regions as much as possible; so it not only handles better intensity inhomogeneity, but also allows more robustness to noise and more flexible initialization in comparison to the original global region and regional-scalable based models.

View Article: PubMed Central - PubMed

Affiliation: Key Laboratory of Optoelectronic Technology and Systems of Ministry of Education, College of Optoelectronic Engineering, Chongqing University, Chongqing 400044, China.

ABSTRACT
We propose a novel region-based geometric active contour model that uses region-scalable discriminant and fitting energy functional for handling the intensity inhomogeneity and weak boundary problems in medical image segmentation. The region-scalable discriminant and fitting energy functional is defined to capture the image intensity characteristics in local and global regions for driving the evolution of active contour. The discriminant term in the model aims at separating background and foreground in scalable regions while the fitting term tends to fit the intensity in these regions. This model is then transformed into a variational level set formulation with a level set regularization term for accurate computation. The new model utilizes intensity information in the local and global regions as much as possible; so it not only handles better intensity inhomogeneity, but also allows more robustness to noise and more flexible initialization in comparison to the original global region and regional-scalable based models. Experimental results for synthetic and real medical image segmentation show the advantages of the proposed method in terms of accuracy and robustness.

Show MeSH

Related in: MedlinePlus

Segmentation results on images without and with intensity inhomogeneity of CV (a), RSF (b), LGIF (c), and our method (d) with different initial contours. The initial contours and the final contours are plotted as green contours and red contours, respectively (color online).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4109108&req=5

fig1: Segmentation results on images without and with intensity inhomogeneity of CV (a), RSF (b), LGIF (c), and our method (d) with different initial contours. The initial contours and the final contours are plotted as green contours and red contours, respectively (color online).

Mentions: We firstly evaluate our method in segmenting a publicly available synthetic image with intensity inhomogeneity and compare it with CV, RSF, and LGIF methods. The searching ranges of the parameters are as follows: w = 0.95 ~ 0.99; k = 0.1~0.3; ν = 0.001∗2552 ~ 0.005∗2552; Δt = 0.1; and μ = 0.1 ~ 0.2/Δt. The original images marked with the segmentation results with different initial contours are reported in Figure 1. From where it is seen when there is a clear object in the slight obscure background, all the methods can capture right object (e.g., images in the first row). When the objects and background are, respectively, homogeneous but the intensities vary among the objects, the local and combined methods can localize the objects but the global method failed (e.g., images in the second row). Furthermore, when there are objects corrupted by the strong nonuniform noise, only our proposed method extracts the desirable objects. This experiment validates the merits of locally and globally combined method. Due to the global minimization, the segmentation results in the first column by CV model are very similar with respect to different initial contours. However, only the ellipse-like object is correctly segmented. In contrast, the RSF model can successfully segment the ellipse-like and rectangle-like objects due to its capability in capturing the local intensity. However, it fails to extract the star-like object that is full much stronger nonuniform noise than others. In addition, the RSF model is sensitive to the initial contours. The LGIF model outperforms RSF model in its robustness on different initial contours. Unfortunately, it still fails to segment the star-like object. As a comparison, our method successfully extracts three objects under different initial contours; meanwhile, it obtains very similar final contours in these testings due to its global optimization.


New region-scalable discriminant and fitting energy functional for driving geometric active contours in medical image segmentation.

Wang X, Niu Y, Tan L, Zhang SX - Comput Math Methods Med (2014)

Segmentation results on images without and with intensity inhomogeneity of CV (a), RSF (b), LGIF (c), and our method (d) with different initial contours. The initial contours and the final contours are plotted as green contours and red contours, respectively (color online).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4109108&req=5

fig1: Segmentation results on images without and with intensity inhomogeneity of CV (a), RSF (b), LGIF (c), and our method (d) with different initial contours. The initial contours and the final contours are plotted as green contours and red contours, respectively (color online).
Mentions: We firstly evaluate our method in segmenting a publicly available synthetic image with intensity inhomogeneity and compare it with CV, RSF, and LGIF methods. The searching ranges of the parameters are as follows: w = 0.95 ~ 0.99; k = 0.1~0.3; ν = 0.001∗2552 ~ 0.005∗2552; Δt = 0.1; and μ = 0.1 ~ 0.2/Δt. The original images marked with the segmentation results with different initial contours are reported in Figure 1. From where it is seen when there is a clear object in the slight obscure background, all the methods can capture right object (e.g., images in the first row). When the objects and background are, respectively, homogeneous but the intensities vary among the objects, the local and combined methods can localize the objects but the global method failed (e.g., images in the second row). Furthermore, when there are objects corrupted by the strong nonuniform noise, only our proposed method extracts the desirable objects. This experiment validates the merits of locally and globally combined method. Due to the global minimization, the segmentation results in the first column by CV model are very similar with respect to different initial contours. However, only the ellipse-like object is correctly segmented. In contrast, the RSF model can successfully segment the ellipse-like and rectangle-like objects due to its capability in capturing the local intensity. However, it fails to extract the star-like object that is full much stronger nonuniform noise than others. In addition, the RSF model is sensitive to the initial contours. The LGIF model outperforms RSF model in its robustness on different initial contours. Unfortunately, it still fails to segment the star-like object. As a comparison, our method successfully extracts three objects under different initial contours; meanwhile, it obtains very similar final contours in these testings due to its global optimization.

Bottom Line: The region-scalable discriminant and fitting energy functional is defined to capture the image intensity characteristics in local and global regions for driving the evolution of active contour.The discriminant term in the model aims at separating background and foreground in scalable regions while the fitting term tends to fit the intensity in these regions.The new model utilizes intensity information in the local and global regions as much as possible; so it not only handles better intensity inhomogeneity, but also allows more robustness to noise and more flexible initialization in comparison to the original global region and regional-scalable based models.

View Article: PubMed Central - PubMed

Affiliation: Key Laboratory of Optoelectronic Technology and Systems of Ministry of Education, College of Optoelectronic Engineering, Chongqing University, Chongqing 400044, China.

ABSTRACT
We propose a novel region-based geometric active contour model that uses region-scalable discriminant and fitting energy functional for handling the intensity inhomogeneity and weak boundary problems in medical image segmentation. The region-scalable discriminant and fitting energy functional is defined to capture the image intensity characteristics in local and global regions for driving the evolution of active contour. The discriminant term in the model aims at separating background and foreground in scalable regions while the fitting term tends to fit the intensity in these regions. This model is then transformed into a variational level set formulation with a level set regularization term for accurate computation. The new model utilizes intensity information in the local and global regions as much as possible; so it not only handles better intensity inhomogeneity, but also allows more robustness to noise and more flexible initialization in comparison to the original global region and regional-scalable based models. Experimental results for synthetic and real medical image segmentation show the advantages of the proposed method in terms of accuracy and robustness.

Show MeSH
Related in: MedlinePlus