A phantom study on component segregation for MR images using ICA.

RATIONALE AND OBJECTIVES A phantom set was devised to evaluate capability of independent component analysis (ICA) as an image filter for magnetic resonance (MR) images to segregate components. MATERIALS AND METHODS Four components (free water [FW], olive oil [OL], 2% and 4% agar gels [2A and 4A, respectively]) were arranged in a phantom set. Seven MR images were obtained with different echo time and repetition time. ICA was performed on 23 combinations of four components. A segregation rate higher than 70% was defined as effective. RESULTS Four-component segregation was obtained in 5 of 23 combinations. The best result showed a mean of 87% across the four components. For each component, there were 20 of 23 for FW, 22 for OL, 9 for 2A, and 16 for 4A. CONCLUSIONS The results demonstrated ICA works as an image filter and provides new contrast images that unambiguously segregate components in MR images. For practical application, the best performance should be obtained when T(1)W, T(2)W, and proton density images are included in the dataset for ICA.

[1]  Erkki Oja,et al.  Independent Component Analysis , 2001 .

[2]  Jean-François Cardoso,et al.  Equivariant adaptive source separation , 1996, IEEE Trans. Signal Process..

[3]  Yen-Wei Chen,et al.  Segmentation of MR Images Using Independent Component Analysis , 2006, KES.

[4]  Yasuyo Kita,et al.  An Attempt for Coloring Multichannel MR Imaging Data , 2001, IEEE Trans. Vis. Comput. Graph..

[5]  Eric Moulines,et al.  A blind source separation technique using second-order statistics , 1997, IEEE Trans. Signal Process..

[6]  J. Stuart.  PROCEEDINGS - PART II , 1993 .

[7]  A. Antoniadis,et al.  Segmentation of magnetic resonance brain images through discriminant analysis , 2003, Journal of Neuroscience Methods.

[8]  N. Logothetis.  What we can do and what we cannot do with fMRI , 2008, Nature.

[9]  Lakhmi C. Jain,et al.  Knowledge-Based Intelligent Information and Engineering Systems , 2004, Lecture Notes in Computer Science.

[10]  Aapo Hyvärinen,et al.  Fast and robust fixed-point algorithms for independent component analysis , 1999, IEEE Trans. Neural Networks.

[11]  Chein-I Chang,et al.  Band Expansion-Based Over-Complete Independent Component Analysis for Multispectral Processing of Magnetic Resonance Images , 2008, IEEE Transactions on Biomedical Engineering.

[12]  Toshiharu Nakai,et al.  Application of independent component analysis to magnetic resonance imaging for enhancing the contrast of gray and white matter , 2004, NeuroImage.

[13]  Terrence J. Sejnowski,et al.  An Information-Maximization Approach to Blind Separation and Blind Deconvolution , 1995, Neural Computation.

[14]  Shun-ichi Amari,et al.  Natural Gradient Works Efficiently in Learning , 1998, Neural Computation.