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 an image with different noise of CV (a), RSF (b), LGIF (c), and our method (d). The initial and the final zero level sets are plotted as the green and the red contours, respectively. The images in each row are with Gaussian noise characterized by zero-mean and variance 0.01, 0.05, and 0.1, respectively (color online).
© Copyright Policy
Related In: Results  -  Collection

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

fig2: Segmentation results on an image with different noise of CV (a), RSF (b), LGIF (c), and our method (d). The initial and the final zero level sets are plotted as the green and the red contours, respectively. The images in each row are with Gaussian noise characterized by zero-mean and variance 0.01, 0.05, and 0.1, respectively (color online).

Mentions: We further evaluate the ability of our method in segmenting image with different noise levels. As seen in Figure 2, the testing images are built from a publicly available synthetic image with different zero-mean and variance (0.01, 0.05, and 0.1) Gaussian noise. The parameters are settled as w = 0.98; k = 0.3; ν = 0.002∗2552; Δt = 0.1; and μ = 0.2/Δt. The kernel size in Gaussian window or binary window is σ = 4 for local region, and σ is settled enough to cover the whole image as a global region, similar to the above experiments. The length regularization term is fixed for all methods. The results of the compared methods are reported in Figure 2. From where it is seen when the image is less noisy, each method can obtain exact object, and some noise in background are captured as objects for RSF and LGIF. Along with the heavy noise, the CV model can extract object from noisy background but the contour seems less accurate, while the RSF and the LGIF models still produce missegmentation due to their local fitting property. In contrast, our method consistently captures the right boundary of the object. Although our method also employs the local fitting energy in a similar way, it additionally incorporates the region-scalable mean discriminant energy functional that enlarges the difference of object and background; so it produces the desirable segmenting contours in this case.


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 an image with different noise of CV (a), RSF (b), LGIF (c), and our method (d). The initial and the final zero level sets are plotted as the green and the red contours, respectively. The images in each row are with Gaussian noise characterized by zero-mean and variance 0.01, 0.05, and 0.1, respectively (color online).
© Copyright Policy
Related In: Results  -  Collection

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

fig2: Segmentation results on an image with different noise of CV (a), RSF (b), LGIF (c), and our method (d). The initial and the final zero level sets are plotted as the green and the red contours, respectively. The images in each row are with Gaussian noise characterized by zero-mean and variance 0.01, 0.05, and 0.1, respectively (color online).
Mentions: We further evaluate the ability of our method in segmenting image with different noise levels. As seen in Figure 2, the testing images are built from a publicly available synthetic image with different zero-mean and variance (0.01, 0.05, and 0.1) Gaussian noise. The parameters are settled as w = 0.98; k = 0.3; ν = 0.002∗2552; Δt = 0.1; and μ = 0.2/Δt. The kernel size in Gaussian window or binary window is σ = 4 for local region, and σ is settled enough to cover the whole image as a global region, similar to the above experiments. The length regularization term is fixed for all methods. The results of the compared methods are reported in Figure 2. From where it is seen when the image is less noisy, each method can obtain exact object, and some noise in background are captured as objects for RSF and LGIF. Along with the heavy noise, the CV model can extract object from noisy background but the contour seems less accurate, while the RSF and the LGIF models still produce missegmentation due to their local fitting property. In contrast, our method consistently captures the right boundary of the object. Although our method also employs the local fitting energy in a similar way, it additionally incorporates the region-scalable mean discriminant energy functional that enlarges the difference of object and background; so it produces the desirable segmenting contours in this case.

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