Limits...
An active contour model for the segmentation of images with intensity inhomogeneities and bias field estimation.

Huang C, Zeng L - PLoS ONE (2015)

Bottom Line: The proposed model first appeared as a two-phase model and then extended to a multi-phase one.The experimental results demonstrate the advantages of our model in terms of accuracy and insensitivity to the location of the initial contours.In particular, our method has been applied to various synthetic and real images with desirable results.

View Article: PubMed Central - PubMed

Affiliation: Key Laboratory of Optoelectronic Technology and System of the Education Ministry of China, Chongqing University, Chongqing, 400044, China; Engineering Research Center of Industrial Computed Tomography Nondestructive Testing of the Education Ministry of China, Chongqing University, Chongqing, 400044, China.

ABSTRACT
Intensity inhomogeneity causes many difficulties in image segmentation and the understanding of magnetic resonance (MR) images. Bias correction is an important method for addressing the intensity inhomogeneity of MR images before quantitative analysis. In this paper, a modified model is developed for segmenting images with intensity inhomogeneity and estimating the bias field simultaneously. In the modified model, a clustering criterion energy function is defined by considering the difference between the measured image and estimated image in local region. By using this difference in local region, the modified method can obtain accurate segmentation results and an accurate estimation of the bias field. The energy function is incorporated into a level set formulation with a level set regularization term, and the energy minimization is conducted by a level set evolution process. The proposed model first appeared as a two-phase model and then extended to a multi-phase one. The experimental results demonstrate the advantages of our model in terms of accuracy and insensitivity to the location of the initial contours. In particular, our method has been applied to various synthetic and real images with desirable results.

No MeSH data available.


Comparisons of Li’s model, LSACM and our model for MR brain image.Column 1 is the original image with red and blue initial contours. Column 2 to 4 is the final segmentation results, the corrected images and the estimated bias field images, respectively. Row 1 to 3 is the results of Li’s model, LSACM and our model, respectively.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4383562&req=5

pone.0120399.g011: Comparisons of Li’s model, LSACM and our model for MR brain image.Column 1 is the original image with red and blue initial contours. Column 2 to 4 is the final segmentation results, the corrected images and the estimated bias field images, respectively. Row 1 to 3 is the results of Li’s model, LSACM and our model, respectively.

Mentions: In the following experiments, we compare our model with Li’s model [29] and LSACM [30] (the code was downloaded from [45]) in the performance of multi-phase MR images. Fig. 11 shows the segmentation and bias-correction results on 3T MR image (from [45], the image size is 174×238), which contains white matter (WM), gray matter (GM), cerebrospinal fluid (CSF) and background (CSF usually as the background in our method). We use red contour to represent ϕ1 = 0, and blue to represent ϕ2 = 0. The first column shows the original image and initial contours, the second column shows the segmentation results, the third and fourth columns show the corrected images and bias fields, respectively. The first, second and third rows show the results of Li’s method, LSACM and our method, respectively. As we see from Fig. 11, our method and LSACM capture more CSF than Li’s method and our method obtain more accuracy of GM (WM) in the center of image than other two methods. The corrected images of LSACM and our method seem similar, which are better than Li’s method. For the brain MR image in Fig. 12, the original image (the image size is 141×202) and initial contours are shown in column 1, row 1, 2 and 3 are the results of Li’s method, LSACM and our method, respectively. LSACM can obtain the accurate boundaries of WG (blue contours), but the segmentation of CSF is unexpected (red contours), which can not be well separated the WM and GM. Li’s method can not segment the GM in the center of image. Fig. 11 and Fig. 12 show that our method have more capacity of WM and GM segmentation. The corrected images of Li’s method and our method seem similar and more vivid than LSACM.


An active contour model for the segmentation of images with intensity inhomogeneities and bias field estimation.

Huang C, Zeng L - PLoS ONE (2015)

Comparisons of Li’s model, LSACM and our model for MR brain image.Column 1 is the original image with red and blue initial contours. Column 2 to 4 is the final segmentation results, the corrected images and the estimated bias field images, respectively. Row 1 to 3 is the results of Li’s model, LSACM and our model, respectively.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4383562&req=5

pone.0120399.g011: Comparisons of Li’s model, LSACM and our model for MR brain image.Column 1 is the original image with red and blue initial contours. Column 2 to 4 is the final segmentation results, the corrected images and the estimated bias field images, respectively. Row 1 to 3 is the results of Li’s model, LSACM and our model, respectively.
Mentions: In the following experiments, we compare our model with Li’s model [29] and LSACM [30] (the code was downloaded from [45]) in the performance of multi-phase MR images. Fig. 11 shows the segmentation and bias-correction results on 3T MR image (from [45], the image size is 174×238), which contains white matter (WM), gray matter (GM), cerebrospinal fluid (CSF) and background (CSF usually as the background in our method). We use red contour to represent ϕ1 = 0, and blue to represent ϕ2 = 0. The first column shows the original image and initial contours, the second column shows the segmentation results, the third and fourth columns show the corrected images and bias fields, respectively. The first, second and third rows show the results of Li’s method, LSACM and our method, respectively. As we see from Fig. 11, our method and LSACM capture more CSF than Li’s method and our method obtain more accuracy of GM (WM) in the center of image than other two methods. The corrected images of LSACM and our method seem similar, which are better than Li’s method. For the brain MR image in Fig. 12, the original image (the image size is 141×202) and initial contours are shown in column 1, row 1, 2 and 3 are the results of Li’s method, LSACM and our method, respectively. LSACM can obtain the accurate boundaries of WG (blue contours), but the segmentation of CSF is unexpected (red contours), which can not be well separated the WM and GM. Li’s method can not segment the GM in the center of image. Fig. 11 and Fig. 12 show that our method have more capacity of WM and GM segmentation. The corrected images of Li’s method and our method seem similar and more vivid than LSACM.

Bottom Line: The proposed model first appeared as a two-phase model and then extended to a multi-phase one.The experimental results demonstrate the advantages of our model in terms of accuracy and insensitivity to the location of the initial contours.In particular, our method has been applied to various synthetic and real images with desirable results.

View Article: PubMed Central - PubMed

Affiliation: Key Laboratory of Optoelectronic Technology and System of the Education Ministry of China, Chongqing University, Chongqing, 400044, China; Engineering Research Center of Industrial Computed Tomography Nondestructive Testing of the Education Ministry of China, Chongqing University, Chongqing, 400044, China.

ABSTRACT
Intensity inhomogeneity causes many difficulties in image segmentation and the understanding of magnetic resonance (MR) images. Bias correction is an important method for addressing the intensity inhomogeneity of MR images before quantitative analysis. In this paper, a modified model is developed for segmenting images with intensity inhomogeneity and estimating the bias field simultaneously. In the modified model, a clustering criterion energy function is defined by considering the difference between the measured image and estimated image in local region. By using this difference in local region, the modified method can obtain accurate segmentation results and an accurate estimation of the bias field. The energy function is incorporated into a level set formulation with a level set regularization term, and the energy minimization is conducted by a level set evolution process. The proposed model first appeared as a two-phase model and then extended to a multi-phase one. The experimental results demonstrate the advantages of our model in terms of accuracy and insensitivity to the location of the initial contours. In particular, our method has been applied to various synthetic and real images with desirable results.

No MeSH data available.