Utilizing prior information for bias field estimation and segmentation of MRI data

The method of utilizing available prior information in the popular FCM algorithm and assesses its benefits in estimating the intensity inhomogeneities and segmenting human brain MRI volumes is studied in this paper. The intensity inhomogeneities in medical images are associated with the acquisition...

Full description

Saved in:
Bibliographic Details
Published in2009 International Conference on Computer Engineering & Systems pp. 327 - 332
Main Authors El-Melegy, M., Mokhtar, H.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.12.2009
Subjects
bias field
Biomedical imaging
Coils
Fuzzy c-mean
Humans
Image segmentation
Intensity inhomogeneity
Iterative algorithms
Magnetic resonance imaging
Medical diagnostic imaging
medical image segmentation
Nonuniform electric fields
Positron emission tomography
Radio frequency
Online AccessGet full text

Cover

More Information
Summary:The method of utilizing available prior information in the popular FCM algorithm and assesses its benefits in estimating the intensity inhomogeneities and segmenting human brain MRI volumes is studied in this paper. The intensity inhomogeneities in medical images are associated with the acquisition sequences and imperfections in the radio-frequency coils in MRI scanners. Presence of intensity inhomogeneities in medical images produces a shading artifact which biases the true voxel intensity. The proposed method modifies the objective function of the standard FCM to take into account any available information about the class centers, and class's pixels distribution throughout the MRI volume. The experiments using 3D synthetic phantoms and real MRI volumes show that the proposed method has considerable better segmentation accuracy, robustness against noise, and needs a smaller number of iterations to reach convergence compared with other most famous reported techniques.
ISBN:1424458420
9781424458424
DOI:10.1109/ICCES.2009.5383246