Constrained optimization of nonparametric entropy-based segmentation of brain structures

Asl, Akbar Abdollahi; Soltanian‐Zadeh, Hamid

doi:10.1109/isbi.2008.4540927

Cited by 6 publications

(8 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our results were compared with the methods proposed by [59], [60], and [61] = 0.76. Virtually all the results indicated that COSA performs better than all the proposed methods.…”

Section: Atlases Selection Using Atrophy Measurementioning

confidence: 99%

Parcellation of Brain Magnetic Resonance Images using Improved Atlases Selection Model

Owate,

Aribisala,

Uwadia

et al. 2023

JRRS

View full text Add to dashboard Cite

Introduction: Multiple atlas-based parcellation model has been demonstrated to perform better than single atlas-based parcellation model in terms of accuracy of the parcellation of human brain Magnetic Resonance Images (MRI). The weakness of the existing multiple atlas-based parcellation models is that the level of accuracy is limited if used for the ageing brain due to the presence of age-related changes such as atrophy. Aim: The aim of this study is to develop a novel multiple atlases selection model that ensures improved accuracy for the parcellation of the ageing brain by combining Cost function with the Similarity metric and Atrophy measure for atlases selection. This model is called COSA. Materials and Methods: A dataset with ten brain MRIs and ten atlases was used. A brain MRI was used one at a time as the target image while the remaining images constituted the source images. Using each target image, consensus atlases were obtained for COSA from the combination of a cost function, similarity index, and atrophy measure. These atlases were consequently used to parcellate the target image. Performance was assessed using Dice Coefficient and COSA was compared with existing atlases selection models. The existing atlases selection models investigated were Normalized Mutual Information (NMI), Mutual Information (MI), Correlation Ratio (CR), Normalized Correlation Ratio (NCR), and Least Square Error (LSE). Results: Mean of Dice Coefficient were: COSA = 0.7495196, NMI = 0.7479508, MI = 0.7473333, Jaccard Index = 0.7392522, CR = 0.7358384, NC = 0.7358043, Atrophy Measure = 0.7300867, LSE = 0.7299367, Single Atlas =0.6830223. Conclusion: Results show that COSA performs better than the existing multiple Atlas-based models.

show abstract

“…Our results were compared with the methods proposed by [59], [60], and [61] = 0.76. Virtually all the results indicated that COSA performs better than all the proposed methods.…”

Section: Atlases Selection Using Atrophy Measurementioning

confidence: 99%

Parcellation of Brain Magnetic Resonance Images using Improved Atlases Selection Model

Owate,

Aribisala,

Uwadia

et al. 2023

JRRS

View full text Add to dashboard Cite

show abstract

“…Several shape representation methods are used in literature. [24][25][26][27][28][29] Cootes et al 27 introduced a point based method. Hong et al 28 used kernel integrals for shape representation and Pizer et al 29 used medial shape representation.…”

Section: Shape Modelmentioning

confidence: 99%

“…In addition, they have initialization problems and heavily depend on the training datasets. Akhondi-Asl and Soltanian-Zadeh 26 introduced a constrained optimization strategy to develop a robust segmentation method. Albeit its superiority in performance, it takes considerable time to execute and is somewhat sensitive to the parameters of the constraints.…”

Section: Figmentioning

confidence: 99%

Two‐stage multishape segmentation of brain structures using image intensity, tissue type, and location informationa)

Akhondi-Asl

Soltanian-Zadeh²

2010

Medical Physics

Self Cite

View full text Add to dashboard Cite

Purpose:The authors propose a fast, robust, nonparametric, entropy-based, coupled, multishape approach to segment subcortical brain structures from magnetic resonance images ͑MRIs͒. Methods: The proposed method uses three types of information: Image intensity, tissue types, and locations of structures. The image intensity information is captured by estimating the probability density function ͑pdf͒ of the image intensities in each structure. The tissue type information is captured by applying an unsupervised tissue segmentation method to the image and estimating a probability mass function ͑pmf͒ for the tissue type of each structure. The location information is captured by estimating pdf of the location of each structure from the training datasets. The resulting pmf's and pdf's are used to define an entropy function whose minimum corresponds to a desirable segmentation of the structures. The authors propose a three-step optimization strategy for the segmentation method. In the first step, a powerful automatic initialization method is developed based on tissue type and location information of the structures. In the second step, a quasi-Newton method is used to optimize the parameters of the energy function. To speed up the iterations, derivatives of the energy function with respect to its parameters are analytically derived and used in the optimization process. In the last step, the limitations related to the prior shape model are removed and a level-set method is applied for the fine tuning of the segmentation results. Results: The proposed method is applied to two different datasets and the results are compared to those of previous methods in literature. Experimental results are presented for lateral ventricles, caudate, thalamus, putamen, pallidum, hippocampus, and amygdala. Conclusions:The results illustrate superior performance of the proposed segmentation method compared to other methods in literature. The execution time of the algorithm is a few minutes, suitable for a variety of applications.

show abstract

“…In [2][3] we have introduced a coupled structures segmentation algorithm that is based on a principal component analysis (PCA) designed to extract shape relationships among structures. In [4] we have added constraint to the shape parameters to achieve a more robust segmentation algorithm. In [5] we have modified the energy function by considering tissue type and location of the structures as independent information.…”

Section: Introductionmentioning

confidence: 99%

Cortical brain structures segmentation using constrained optimization and intensity coupling

Akhondi-Asl

Warfield

2011

2011 IEEE International Symposium on Biomedical Imaging: From Nano to Macro

View full text Add to dashboard Cite

Brain image segmentation is one of the most important applications in medicine and also is one of the most challenging topics in the field of medical image processing. In general, most automatic segmentation methods consist of an energy function, a shape model, and an optimization strategy. Each plays an important role in the design of an accurate segmentation algorithm. Here we introduce a modified version of a coupled structure segmentation algorithm that is based on earlier paper. Specifically, we have 1) utilized a multiple atlas strategy to estimate a joint probability mass function of the location and tissue type information of the structures; 2) analyzed the relationship among the various structures to achieve more robust probability density function (pdf) estimation; 3) added a constraint to the optimization process to minimize intersection among the different structures; and 4) demonstrated the effectiveness of the method for the segmentation of certain brain structures.

show abstract

Constrained optimization of nonparametric entropy-based segmentation of brain structures

Cited by 6 publications

References 17 publications

Parcellation of Brain Magnetic Resonance Images using Improved Atlases Selection Model

Parcellation of Brain Magnetic Resonance Images using Improved Atlases Selection Model

Two‐stage multishape segmentation of brain structures using image intensity, tissue type, and location informationa)

Cortical brain structures segmentation using constrained optimization and intensity coupling

Contact Info

Product

Resources

About