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 MR brain image with intensity inhomogeneity by CV (a), RSF (b), LGIF (c) and our method (d) with different initial contours. The initial and the final zero level sets are plotted as the green and the red contours, respectively (color online).
© Copyright Policy
Related In: Results  -  Collection

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

fig5: Segmentation results on an MR brain image with intensity inhomogeneity by CV (a), RSF (b), LGIF (c) and our method (d) with different initial contours. The initial and the final zero level sets are plotted as the green and the red contours, respectively (color online).

Mentions: We further evaluate our method in segmenting MRI brain images with intensity inhomogeneity. These images usually not only contain weak boundaries between gray matter and white matter due to low contrast and partial volume effect, but also present intensity inhomogeneity arisen from the acquisition style. The parameters are searched in the range w = 0.1~0.99; k = 0.3; ν = 0.001∗2552 ~ 0.005∗2552; Δt = 0.1; and μ = 0.2/Δt for an optimal segmentation of each method. Figures 5 and 6 present the performance of our method and compared methods on segmenting two MR brain images with remarkable intensity inhomogeneity. The intensities of white matter in Figure 5 have visible intensity variations and are obscure in most cortex regions. In this experiment, Both CV and RSF models cannot segment the details of the white matter accurately. The results of LGIF are slightly better than those of CV and RSF models, but there are still some segmentation errors, especially in different initial contours. Our method outperforms them in successfully segmenting the white matter in the region with nonuniform intensities. The white matter object in Figure 6 is not corrupted by the intensity inhomogeneity but is more complex in the cortex region. In this experiment, it is observed that our model yields more accurate results than other methods. The experimental results again illustrate the merit of our proposed method: the abilities to deal with intensity inhomogeneity, weak boundaries and complex background, and robustness to noise. These MR images are rather noisy and the object boundary is very weak. As a combined model, LGIF performs better than CV and RSF models; however, it is still easy to misclassify the gray matter as white matter. Again, this fact verifies the limitation of simple combination of local information and global information to drive the curve evolution. On the contrary, the mean discriminant energy functional in our model plays more important role in driving the curve evolution rightly.


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 MR brain image with intensity inhomogeneity by CV (a), RSF (b), LGIF (c) and our method (d) with different initial contours. The initial and the final zero level sets are plotted as the green and the red contours, respectively (color online).
© Copyright Policy
Related In: Results  -  Collection

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

fig5: Segmentation results on an MR brain image with intensity inhomogeneity by CV (a), RSF (b), LGIF (c) and our method (d) with different initial contours. The initial and the final zero level sets are plotted as the green and the red contours, respectively (color online).
Mentions: We further evaluate our method in segmenting MRI brain images with intensity inhomogeneity. These images usually not only contain weak boundaries between gray matter and white matter due to low contrast and partial volume effect, but also present intensity inhomogeneity arisen from the acquisition style. The parameters are searched in the range w = 0.1~0.99; k = 0.3; ν = 0.001∗2552 ~ 0.005∗2552; Δt = 0.1; and μ = 0.2/Δt for an optimal segmentation of each method. Figures 5 and 6 present the performance of our method and compared methods on segmenting two MR brain images with remarkable intensity inhomogeneity. The intensities of white matter in Figure 5 have visible intensity variations and are obscure in most cortex regions. In this experiment, Both CV and RSF models cannot segment the details of the white matter accurately. The results of LGIF are slightly better than those of CV and RSF models, but there are still some segmentation errors, especially in different initial contours. Our method outperforms them in successfully segmenting the white matter in the region with nonuniform intensities. The white matter object in Figure 6 is not corrupted by the intensity inhomogeneity but is more complex in the cortex region. In this experiment, it is observed that our model yields more accurate results than other methods. The experimental results again illustrate the merit of our proposed method: the abilities to deal with intensity inhomogeneity, weak boundaries and complex background, and robustness to noise. These MR images are rather noisy and the object boundary is very weak. As a combined model, LGIF performs better than CV and RSF models; however, it is still easy to misclassify the gray matter as white matter. Again, this fact verifies the limitation of simple combination of local information and global information to drive the curve evolution. On the contrary, the mean discriminant energy functional in our model plays more important role in driving the curve evolution rightly.

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