#### DMCA

## A review on MR image intensity inhomogeneity correction (2006)

Venue: | International Journal of Biomedical Imaging |

Citations: | 21 - 0 self |

### Citations

5116 | Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images
- Geman, Geman
- 1984
(Show Context)
Citation Context ...n the IIH modeling, one would likely arrive at a smoother approximation of the IIH map. A useful tool for spatial modeling is the Markov random field (MRF), which is first employed by Geman and Geman =-=[72]-=- for image segmentation. Held et al. [69] haveemployed the MRF to model IIH. According to the HamerslyClifford theorem [73], the prior probability p(y) isgivenby the Gibbs distribution p(y) ∝ exp { − ... |

1020 | Non-Uniform Random Variate Generation
- Devroye
- 1986
(Show Context)
Citation Context ...econvolution problem by realizing that if v1 and v2 are two independent random variables with distributions V1 and V2, respectively, then the distribution of their sum is the convolution of V1 and V2 =-=[75]-=-. To constrain the solution space, the IIH field is modeled as a Gaussian distribution with small variance. Code for this method is publicly available. 1 5. OTHER APPROACHES 5.1. Comparison between lo... |

639 | Segmentation of brain MR images through a hidden Markov random field model and the expectation maximization algorithm. - Zhang, Brady, et al. - 2001 |

409 |
A nonparametric method for automatic correction of intensity nonuniformity
- Sled, Zijdenbos, et al.
- 1998
(Show Context)
Citation Context ... y j + β y 〈i,j〉 i 2 i , (12) where 〈i, j〉 sums over every voxel i and its neighbours j. 4.3. The N3 method Different from most IIH correction method which involves a classification step, Sled et al. =-=[74]-=- proposed a nonparametric nonuniform intensity normalization (N3) method which searches for the IIH field to maximize the frequency content of the image intensity distribution. The method simplified t... |

392 |
Spatial registration and normalization of images.
- riston, Ashburner, et al.
- 1995
(Show Context)
Citation Context ...me-consuming, hence techniques such as multigrid computing may help reduce the computation load. 5.4. Registration against template Image registration has also been utilized to aid the IIH correction =-=[89, 90]-=-. In [90], the patient data was registered against a tissue reference template, which allows to estimate the IIH map by direct comparison between two images. Here human intervention was employed to en... |

323 | Unified segmentation.
- Ashburner, KJ
- 2005
(Show Context)
Citation Context ...ion against a template could greatly help derive a high-quality image. In addition, there even exist attempts to address these three processes, IIH correction, segmentation, and registration together =-=[93, 94]-=-. Among some processes, their relationship may not be very intimate, but different implementation order could result in different performance. In [68], it was found that denoising after IIH correction... |

283 | Brainweb: online interface to a 3D MRI simulated brain database.
- Cocosco, Kollokian, et al.
- 1997
(Show Context)
Citation Context ...is difficulty. One is to approximate the golden standard by experts’ estimation, which is often a tedious task. Alternatively, we can use synthetic data. In IIH correction, the simulated brain images =-=[7, 96]-=- from the Montreal Neurological Institute 3 have been widely employed for validation. (1) Criteria The criteria that have been frequently used are listed in the following. Cr-I The variance of the ful... |

272 |
Design and construction of a realistic digital brain phantom”,
- Collins, Zijdenbos, et al.
- 1998
(Show Context)
Citation Context ...model in describing the IIH effect is x = αx ′ + ξ, (1) where α denotes the IIH effect and ξ the noise. Notation of bold letters refers to 2D or 3D MR data. Figure 2 displays the widely used BrainWeb =-=[7]-=- simulated MR images, where on Figure 2(a) is the original image, on Figure 2(b) the image with IIH artifact, on Figure 2(c) the image with noise, and on Figure 2(d) the image with both IIH artifact a... |

169 | Voxel-based morphometry — the methods - Ashburner, Friston - 2000 |

158 | Adaptive fuzzy segmentation of magnetic resonance images. - Pham, Prince - 1999 |

137 | Magnetic resonance image tissue classification using a partial volume model
- Shattuck, Sandor-Leahy, et al.
- 2001
(Show Context)
Citation Context ...the combined tissue class even though there is no artifact like IIH. A possible solution is to carry out more detail tissue classification in each local region. For example, the technique proposed in =-=[78, 79]-=- 2 estimated the global tissue mean values by empirical thresholding, while the local statistics are derived through fitting the local histogram with a theoretical distribution. After obtaining estima... |

118 | A modified fuzzy c-means algorithm for bias field estimation and segmentation - Ahmed, Yamany, et al. - 2002 |

108 |
Haselgrove C, et al. Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron 2002; 33: 341–55
- Fischl, DH, et al.
(Show Context)
Citation Context ...ion against a template could greatly help derive a high-quality image. In addition, there even exist attempts to address these three processes, IIH correction, segmentation, and registration together =-=[93, 94]-=-. Among some processes, their relationship may not be very intimate, but different implementation order could result in different performance. In [68], it was found that denoising after IIH correction... |

105 | Brainsuite: An automated cortical surface identification tool.
- Shattuck, Leahy
- 1935
(Show Context)
Citation Context ...the combined tissue class even though there is no artifact like IIH. A possible solution is to carry out more detail tissue classification in each local region. For example, the technique proposed in =-=[78, 79]-=- 2 estimated the global tissue mean values by empirical thresholding, while the local statistics are derived through fitting the local histogram with a theoretical distribution. After obtaining estima... |

94 |
Estimating the bias field of MR images,”
- Guillemaud, Brady
- 1997
(Show Context)
Citation Context ...IH, the above method is simply the tissue classification using a mixture Gaussian model. Hence, this method essentially interleaves the IIH correction with a Gaussian classifier. Guillemaud and Brady =-=[68]-=- observed that the effect of IIH correction by Wells et al. algorithm is substantially affected by the Gaussian classifier. In real images, it is very possible for the histogram to deviate from the mi... |

87 | Markov random field segmentation of brain MR images",
- Held, Kops, et al.
- 1997
(Show Context)
Citation Context ...ve at a smoother approximation of the IIH map. A useful tool for spatial modeling is the Markov random field (MRF), which is first employed by Geman and Geman [72] for image segmentation. Held et al. =-=[69]-=- haveemployed the MRF to model IIH. According to the HamerslyClifford theorem [73], the prior probability p(y) isgivenby the Gibbs distribution p(y) ∝ exp { − U(y) } (11) for a tissue and (2) that the... |

71 | Parametric estimate of intensity inhomogeneities applied to MRI,"
- Styner, Brechbuhler, et al.
- 2000
(Show Context)
Citation Context ...cantly for the pixels of the same tissue and the intensity values overlap markedly between the pixels of the different tissues. For comparison, the IIH corrected images by a surface fitting technique =-=[6]-=-aregiveninFigure 1(b),fromwhich the improvement in image quality is clearly visible. The estimated IIH maps are given in Figure 1(c). Let x denote the measured intensity and x ′ the true intensity. Th... |

65 | An adaptative fuzzy C-means algorithm for image segmentation in the presence of intensity inhomogeneities - Pham, Prince - 1999 |

62 | Multimodal image coregistration and partitioning — a unified framework. - Ashburner, Friston - 1997 |

61 |
Segmentation of multiple sclerosis lesions in intensity corrected multispectral MRI
- Johnston, Aitkins, et al.
- 1996
(Show Context)
Citation Context ...[68]. Overall, the N3 method performs the most stable for all simulated images. Velthuizen et al. [29] have evaluated four IIH correction methods (a phantom method [17], two lowpass filtering methods =-=[36, 39]-=-, and a surface fitting method with reference points selected from white matter [48]) in brain tumor segmentation. The surface fitting method was found to be inferior to others, which could be due to ... |

57 |
Adaptive segmentation of MRI data,”
- Crimson, Kikinis, et al.
- 1996
(Show Context)
Citation Context ...log-transform of x) as follows: β = max β p( β | y ) . (4) This is called the maximum a posterior (MAP) estimate and, by the Bayes rule, is equivalent to β = max β p( y | β ) p(β). (5) Wells et al. =-=[67]-=- used the Gaussian distribution to model the entire log-transformed bias field and the observed intensity at voxel i: p(β) = Gψβ(β), (6) p ( ) ( yi | Γi, βi = GψΓ yi − μ i ( ) ) Γi − βi , (7) where Γi... |

57 | An accurate and efficient Bayesian method for automatic segmentation of brain - Marroquín, Vemuri, et al. - 2002 |

56 |
Scale-based fuzzy connected image segmentation: Theory, algorithms, and validation,”
- Saha, Udupa, et al.
- 2000
(Show Context)
Citation Context ...lgorithm. Liew and Yan [58] approximated the IIH as a stack of Bspline surfaces with continuity constraints across slices. The estimation of IIH interwines with a fuzzy c-means clustering process. In =-=[61, 62]-=-, segmentation that utilizes local scale as homogeneous criteria has been presented and applied to IIH correction. When a statistical classifier, such as Gaussian classifier or random field modeling, ... |

55 |
Retrospective correction of MR intensity inhomogeneity by information minimization
- Likar, Viergever, et al.
- 2001
(Show Context)
Citation Context ...avoid this pitfall, constraints over the solution space are necessary. Mangin [55] constrained the solution to minimize the distance between the mean values of the restored and the original image. In =-=[57]-=-, the restored image was constrained to have the same mean value as the original one. Evidently, other quantities relevant to image features, variance for example, can also be applicable in a similar ... |

53 |
Correction of intensity variations in MR images for computeraided tissue classification,"
- Dawant, Zijidenbos, et al.
- 1993
(Show Context)
Citation Context ...voxels SI ⊆ I which convey information about the IIH map. Among these methods, the essential difference lies in the identification of SI, and hence decoding the IIH information from SI. Dawant et al. =-=[48]-=- proposed to manually select SI such that they belong to the same type of tissue. As a result, the intensity variation among these voxels can largely be attributed to IIH. However, expert supervision ... |

49 | Entropy minimization for automatic correction of intensity nonuniformity.
- Mangin
- 2004
(Show Context)
Citation Context ...ved through searching the parameter space of the IIH model such that the entropy of the image is reduced. It should be pointed out that direct minimization on the entropy would lead to the null field =-=[55, 66]-=-. To avoid this pitfall, constraints over the solution space are necessary. Mangin [55] constrained the solution to minimize the distance between the mean values of the restored and the original image... |

47 | Brain development, gender and IQ in children. A volumetric imaging study. - Reiss, Abrams, et al. - 1996 |

45 | An adaptive spatial fuzzy clustering algorithm for 3-D MR image segmentation,”
- Liew, Yan
- 2003
(Show Context)
Citation Context ...ity is poor due to perturbation such as noise, partial volume artifact, or IIH. Beside that, not every clinical department can afford the computer cost to run the parallel LCJ algorithm. Liew and Yan =-=[58]-=- approximated the IIH as a stack of Bspline surfaces with continuity constraints across slices. The estimation of IIH interwines with a fuzzy c-means clustering process. In [61, 62], segmentation that... |

41 | Segmentation of MR brain images into cerebrospinal fluid spaces, white and gray matter. - Lim, Pfefferbaum - 1989 |

40 | A homogeneity correction method for magnetic resonance imaging with time-varying gradients,” - Noll, Meyer, et al. - 1991 |

37 |
Optimized homo-morphic unsharp masking for MR grayscale inhomogeneity correction,"
- Brinkmann, Manduca, et al.
- 1998
(Show Context)
Citation Context ...ages has been carried out in [31]. Due to their simplicity and efficiency in implementation, lowpass filtering methods have been widely used [30–43]. For a summary, interested readers are referred to =-=[42]-=-. Also in [42], the impact of filter width on IIH correction was investigated and it was found that these methods should be used with care to avoid intensity distortion and artificial artifacts in the... |

35 | Segmentation of MR images with intensity inhomogeneities,”
- Rajapakse, Kruggel
- 1998
(Show Context)
Citation Context ...he BrainSuite package (http://neuroimage.usc.edu/).6 International Journal of Biomedical Imaging other parameters. Then one way to estimate β can be obtained by ∂Φ ∂β = 0. (13) Rajapakse and Kruggel =-=[84]-=- used an MRF formulation, whereas Farag and his group exploited the fuzzy c-means clustering framework [85–88]. It is noted that in these methods no explicit assumption has been made on the IIH field,... |

30 |
Retrospective correction of intensity inhomogeneities in MRI,"
- Meyer, Bland, et al.
- 1995
(Show Context)
Citation Context ...er, expert supervision to select points is time consuming and error prone, especially for volume data. Wang et al. [59] haspresentedanautomatedmethod for generating the reference points. Meyer et al. =-=[49]-=- employed the LCJ method [60] topreliminarily segment the image and then fit a smooth functional over the segmented image. The LCJ segmentation method assumes the image to be piecewise smooth and requ... |

25 | A.: A Fast Model Independent Method for Automatic Correction of Intensity Non-uniformity
- Vokurka, Thacker, et al.
- 1999
(Show Context)
Citation Context ...ombined with those calculated in the rows/columns from the initial to the current one. The rationale underlying the computation is obscure from the description. However, it is very similar to that in =-=[81]-=-. In the latter technique, a smooth “variation” image was firstly derived from normalized intensity gradient field, where pixels with low intensity or high gradient magnitude are excluded. Then numeri... |

23 |
Computational Neuroanatomy
- Ashburner
- 2000
(Show Context)
Citation Context ...n that utilizes local scale as homogeneous criteria has been presented and applied to IIH correction. When a statistical classifier, such as Gaussian classifier or random field modeling, is exploited =-=[50, 53, 63]-=-, the process is similar to the parameter estimation in Section 4, where the parameters are associated with a probability distribution and can be estimated with common statistical estimation methods s... |

22 |
Rapid and effective correction of RF inhomogeneity for high field magnetic resonance imaging
- Cohen, DuBois, et al.
- 2000
(Show Context)
Citation Context ...nalysis allows one to select an optimal scale from which the approximate band in the wavelet transform domain is used for estimating the IIH map. In [47], an improvement of a lowpass filtering method =-=[43]-=- was presented. The method varies the filter kernel size to minimize the segmentation error. The idea is generally similar to [45, 46] in addressing IIH correction from the scale space, but differs in... |

17 | Inhomogeneity correction using an estimated linear field map,” - Irarrazabal, Meyer, et al. - 1996 |

17 |
Automated modelbased bias field correction of MR images of the brain
- Leemput, Maes, et al.
- 1999
(Show Context)
Citation Context ...n that utilizes local scale as homogeneous criteria has been presented and applied to IIH correction. When a statistical classifier, such as Gaussian classifier or random field modeling, is exploited =-=[50, 53, 63]-=-, the process is similar to the parameter estimation in Section 4, where the parameters are associated with a probability distribution and can be estimated with common statistical estimation methods s... |

17 | Retrospective shading correction based on entropy minimization - Likar, Viergever, et al. - 2000 |

16 |
Maximum Likelihood Estimation of the Bias Field
- Prima, Ayache, et al.
- 2001
(Show Context)
Citation Context ...n that utilizes local scale as homogeneous criteria has been presented and applied to IIH correction. When a statistical classifier, such as Gaussian classifier or random field modeling, is exploited =-=[50, 53, 63]-=-, the process is similar to the parameter estimation in Section 4, where the parameters are associated with a probability distribution and can be estimated with common statistical estimation methods s... |

16 |
Interplay between intensity standardization and inhomogeneity correction in MR image processing,”
- Madabhushi, Udupa
- 2005
(Show Context)
Citation Context ... may not be very intimate, but different implementation order could result in different performance. In [68], it was found that denoising after IIH correction is more preferable. Madabhushi and Udupa =-=[95]-=- investigated the interplay between IIH correction and intensity standardization, and concluded that the better sequence is IIH correction followed by intensity standardization. 6.2. Validation and co... |

15 | Imaging of the active B1 field - Stollberger, Wach - 1996 |

15 | An algorithm for compensation of surfacecoil images for sensitivity of the surface coil. Magn Reson Imaging - Haselgrove, Prammer - 1986 |

13 | Compensation for surface coil sensitivity variation in magnetic resonance imaging. Magn Reson Imaging - Narayana, Brey, et al. - 1988 |

13 | Methods for measuring brain morphologic features on magnetic resonance images. - Jernigan, Press, et al. - 1990 |

13 |
A wavelet-based approximation of surface coil sensitivity profiles for correction of image intensity inhomogeneity and parallel imaging reconstruction,”
- Lin, Chen, et al.
- 2003
(Show Context)
Citation Context ...ct the true image, after which the ratio between the observed and estimated image is used for IIH map approximation. Recently, lowpass filtering methods have been extended using the wavelet transform =-=[45, 46]-=- and were shown to be effective in removing IIH in images acquired by surface coils and phase array coils. Compared with usual lowpass filtering methods, the multiresolution analysis allows one to sel... |

13 |
Local histogram correction of MRI spatially dependent image pixel intensity nonuniformity
- DeCarli, Murphy, et al.
- 1996
(Show Context)
Citation Context ...itive to the estimation of tissue statistics, which is nontrivial in practice. When GM and WM are combined as one class and the local statistic is estimated from a sample in a local region as done in =-=[76, 77]-=-, the method can be regarded as a generalized white matter method by Dawant et al. [48], where the reference tissue is the combination of GM and WM. Also, it can be taken as a lowpass filtering estima... |

12 | Generalized scale: theory, algorithms, and application to image inhomogeneity correction.
- Madabhushi, Udupa, et al.
- 2006
(Show Context)
Citation Context ...lgorithm. Liew and Yan [58] approximated the IIH as a stack of Bspline surfaces with continuity constraints across slices. The estimation of IIH interwines with a fuzzy c-means clustering process. In =-=[61, 62]-=-, segmentation that utilizes local scale as homogeneous criteria has been presented and applied to IIH correction. When a statistical classifier, such as Gaussian classifier or random field modeling, ... |

12 |
MRI intensity inhomogeneity correction by combining intensity and spatial information,”
- Vovk, Pernuˇs, et al.
- 2004
(Show Context)
Citation Context ...d. Then numerical integration was applied to the “variation” (first-order derivative) image to obtain an image, which only contains small variations and was used to determine the IIH map. Vovk et al. =-=[82, 83]-=- proposed to use the probability distribution of image features for IIH correction, where the image feature includes the intensity and the second spatial derivative of the image. Similar to the usual ... |

11 | Image nonuniformity in magnetic resonance imaging: its magnitude and methods for its correction, The British J. of Radiology 60 - Condon, Patterson, et al. - 1987 |

11 |
A Parallel Technique for Signal-Level Perceptual Organization,”
- Liou, Chiu, et al.
- 1991
(Show Context)
Citation Context ...ect points is time consuming and error prone, especially for volume data. Wang et al. [59] haspresentedanautomatedmethod for generating the reference points. Meyer et al. [49] employed the LCJ method =-=[60]-=- topreliminarily segment the image and then fit a smooth functional over the segmented image. The LCJ segmentation method assumes the image to be piecewise smooth and requires that the different objec... |

9 |
Review and evaluation of MRI nonuniformity corrections for brain tumor response measurements,” Med
- Velthuizen, Heine, et al.
- 1998
(Show Context)
Citation Context ...to lowpass filtering, statistical modeling, surface fitting and others, which are detailed, respectively, in the following sections. For an early literature review, interested readers are referred to =-=[29]-=-, where an evaluation of the IIH correction effect for brain tumor segmentation is also reported. This paper attempts to summarize the recent progress and focus will be on mathematical modeling for II... |

9 |
Comparative Evaluation of Retrospective Shading Correction Methods,"
- Tomazevic, Likar, et al.
- 2002
(Show Context)
Citation Context ...ltering, can be adopted for IIH removal and the converse also holds. An investigation of applying IIH correction methods to deal with the shading problem in microscopic images has been carried out in =-=[31]-=-. Due to their simplicity and efficiency in implementation, lowpass filtering methods have been widely used [30–43]. For a summary, interested readers are referred to [42]. Also in [42], the impact of... |

9 | Intensity correction in surface coil imaging - Axel, Costantini, et al. |

8 |
A new variational shape-from-orientation approach to correcting intensity inhomogeneities in magnetic resonance images
- Lai, Fang
- 1999
(Show Context)
Citation Context ... map by direct comparison between two images. Here human intervention was employed to ensure the correct correspondence between the distorted and the reference image. 5.5. Shape recovery Lai and Fang =-=[91]-=- transformed the IIH correction into the problem of shape recovery with orientation constraint and solved the latter using regularization theory. The approach may result in solving a linear equation w... |

7 | Phase and sensitivity of receiver coils in magnetic resonance imaging - McVeigh, Bronskill, et al. - 1986 |

7 | et al. Generalized reconstruction of phase contrast MRI: analysis and correction of the effect of gradient field distortions. Magn Reson Med 2003 - Markl, Bammer, et al. |

7 | et al. Phased array detectors and an automated intensity-correction algorithm for high-resolution MR imaging of the human brain. Magn Reson Med - LL, Carvajal, et al. |

7 |
et al. Qualitative and Quantitative Evaluation of Six Algorithms for Correcting Intensity Nonuniformity Effects. NeuroImage (2001) 13: 931–934
- Arnold, Liow, et al.
(Show Context)
Citation Context ...ved through searching the parameter space of the IIH model such that the entropy of the image is reduced. It should be pointed out that direct minimization on the entropy would lead to the null field =-=[55, 66]-=-. To avoid this pitfall, constraints over the solution space are necessary. Mangin [55] constrained the solution to minimize the distance between the mean values of the restored and the original image... |

7 |
Accurate template-based correction of brain MRI intensity distortion with application to dementia and aging
- Studholme, Cardenas, et al.
- 2004
(Show Context)
Citation Context ...me-consuming, hence techniques such as multigrid computing may help reduce the computation load. 5.4. Registration against template Image registration has also been utilized to aid the IIH correction =-=[89, 90]-=-. In [90], the patient data was registered against a tissue reference template, which allows to estimate the IIH map by direct comparison between two images. Here human intervention was employed to en... |

6 | Measurement and correction of transmitter and receivers induced nonuniformities in vivo. Magn Reson Med - Wang, Qiu, et al. - 2005 |

6 |
Desco, Method for bias field correction of brain T1-weighted magnetic resonance images minimizing segmentation error
- Gispert, Reig, et al.
(Show Context)
Citation Context ...al lowpass filtering methods, the multiresolution analysis allows one to select an optimal scale from which the approximate band in the wavelet transform domain is used for estimating the IIH map. In =-=[47]-=-, an improvement of a lowpass filtering method [43] was presented. The method varies the filter kernel size to minimize the segmentation error. The idea is generally similar to [45, 46] in addressing ... |

6 |
Post-acquisition correction of MR inhomogeneities
- Lee, Vannier
- 1996
(Show Context)
Citation Context ...itive to the estimation of tissue statistics, which is nontrivial in practice. When GM and WM are combined as one class and the local statistic is estimated from a sample in a local region as done in =-=[76, 77]-=-, the method can be regarded as a generalized white matter method by Dawant et al. [48], where the reference tissue is the combination of GM and WM. Also, it can be taken as a lowpass filtering estima... |

6 |
Tervonen O. Image feature based automatic correction of low-frequency spatial intensity variations in MR images. Magn Reson Imaging
- Koivula, Alakuijala
- 1997
(Show Context)
Citation Context ...ntially relate the IIH correction to tissue segmentation, and the implicit assumptions are (1) the constant intensity 5.2. Image feature-based methods An image feature-based technique was reported in =-=[80]-=-, where the IIH correction was decomposed into row and column correction. The correction factor at a voxel is firstly related to first-order difference at other voxels in the same row/column, then com... |

5 | Radio frequency magnetic field mapping of a 3Tesla birdcage coil: Experimental and theoretical dependence on sample properties - Alecci, Collins, et al. - 2001 |

5 | Narayana, “Correction for intensity falloff in surface coil magnetic resonance imaging - Brey, A - 1988 |

5 | Correction for B(1) and B(0) variations in quantitative T(2) measurements using - Sled, Pike - 2000 |

5 | Optimization of tissue segmentation of brain MR images based on multispectral 3D Feature maps", Magn. Reson. Imaging 17 - Mohamed, Vinitski, et al. - 1999 |

5 |
Effect of radio frequency inhomogeneity correction on the reproducibility of intra-cranial volumes using MR image data. Magnetic Resonance in Medicine
- Narayana, Bourthakour
- 1995
(Show Context)
Citation Context ...[68]. Overall, the N3 method performs the most stable for all simulated images. Velthuizen et al. [29] have evaluated four IIH correction methods (a phantom method [17], two lowpass filtering methods =-=[36, 39]-=-, and a surface fitting method with reference points selected from white matter [48]) in brain tumor segmentation. The surface fitting method was found to be inferior to others, which could be due to ... |

4 | Volume imaging with MR phased arrays - Hayes, Hattes, et al. - 1991 |

4 | Surface coil MR imaging of the human brain with an analytic reception profile correction - Moyher, Vigneron, et al. - 1995 |

4 | Ed.: Intensity Correction of Phased-Array Surface Coil Images - Murakami, Hayes, et al. - 1996 |

4 |
Correction of bias field in MR images using singularity function analysis
- Luo, Zhu, et al.
- 2005
(Show Context)
Citation Context ...or MR images, due to the overlapping spectrum between the patient data and the IIH, the effectiveness of most conventional lowpass filtering in removing the IIH is generally quite limited. Luo et al. =-=[44]-=- presented a technique to recover lowfrequency components which correspond to anatomic structure and are lost during the lowpass filtering. The method expresses the signal with a linear combination of... |

4 |
C.: A Multi-scale Method for Automatic Correction of Intensity Non-Uniformity
- Han, Hatsukami, et al.
- 2001
(Show Context)
Citation Context ...ct the true image, after which the ratio between the observed and estimated image is used for IIH map approximation. Recently, lowpass filtering methods have been extended using the wavelet transform =-=[45, 46]-=- and were shown to be effective in removing IIH in images acquired by surface coils and phase array coils. Compared with usual lowpass filtering methods, the multiresolution analysis allows one to sel... |

4 | A Modified Fuzzy C-Means Algorithm for MRI Bias Field Estimation and Adaptive Segmentation - Ahmed, Yamany, et al. - 1999 |

3 | et al., “Comparison of linear and circular polarization for magnetic resonance imaging - Glover, Hayes, et al. - 1985 |

3 |
Tos, “Correction of intensity nonuniformity
- Wicks, Barker, et al.
- 1993
(Show Context)
Citation Context ...consistent with that pointed out in [68]. Overall, the N3 method performs the most stable for all simulated images. Velthuizen et al. [29] have evaluated four IIH correction methods (a phantom method =-=[17]-=-, two lowpass filtering methods [36, 39], and a surface fitting method with reference points selected from white matter [48]) in brain tumor segmentation. The surface fitting method was found to be in... |

2 | et al., “A unified variational approach to denoising and bias correction - Fan - 2003 |

2 |
MRI sensitivity correction and tissue classification
- Ashburner, Friston
- 1998
(Show Context)
Citation Context ...tation could be less affected by the IIH artifact. A more comprehensive study was presented in [66], where six algorithms, n3 [74], hum [42], eq [43], bfc [78], spm (statistical parametric mapping) 4 =-=[52]-=-, and cma 5 were compared against BrainWeb-simulated data as well as real volumes including repeated scans of the same subject, scans under different magnetic fields and different scanners. Three of t... |

2 |
Mammographic image restoration using maximum entropy deconvolution’, Phys
- Jannetta, Jackson, et al.
- 2004
(Show Context)
Citation Context ...nimization As a frequently used criterion to characterize the intensity distribution of an image, entropy has been employed to design algorithms for image restoration, thresholding, or classification =-=[64, 65]-=-. Also, it has been utilized to quantify the image property with IIH present and guide the parameter searching for IIH removal [54–57]. It is assumed that the intensity distribution of the original im... |

2 |
Spatial interation and statistical analysis of lattice systems
- Besag
- 1974
(Show Context)
Citation Context ...is the Markov random field (MRF), which is first employed by Geman and Geman [72] for image segmentation. Held et al. [69] haveemployed the MRF to model IIH. According to the HamerslyClifford theorem =-=[73]-=-, the prior probability p(y) isgivenby the Gibbs distribution p(y) ∝ exp { − U(y) } (11) for a tissue and (2) that the intensity variation within a tissue is solely due to IIH. Not surprisingly, these... |

2 |
Multi-feature intensity inhomogeneity correction
- Vovk, Pernu, et al.
- 2004
(Show Context)
Citation Context ...d. Then numerical integration was applied to the “variation” (first-order derivative) image to obtain an image, which only contains small variations and was used to determine the IIH map. Vovk et al. =-=[82, 83]-=- proposed to use the probability distribution of image features for IIH correction, where the image feature includes the intensity and the second spatial derivative of the image. Similar to the usual ... |

2 |
Correcting nonuniformities in MRI intensities using entropy minimization based on an elastic model
- Bansal, Staib, et al.
- 2004
(Show Context)
Citation Context ...y with orientation constraint and solved the latter using regularization theory. The approach may result in solving a linear equation with a large matrix. 5.6. Deformed thin plate model Bansal et al. =-=[92]-=- modelled the IIH field as a thin plate deforming elastically under a body force: μ∇ 2 β +(λ + μ)∇(∇·β)+b(β) = 0, (14) where μ and λ are the elasticity constants. The body force b(β) is evaluated to m... |

1 | Correction of off resonance-related distribution in echo-planar imaging using EPI-based field maps - Reber, Wong, et al. - 1998 |

1 | Correction of B1 inhomogeneities using echoplanar imaging of water - Thulborn, Boada, et al. - 1998 |

1 | et al., “Different excitation and reception distributions with a single-loop transmitreceive surface coil near a head-sized spherical phantom at 300 MHz - Collins, Yang, et al. - 2002 |

1 |
Improved version of white matter method for correction of non-uniform intensity in MR images: application to the quantification of rates of brain atrophy in Alzheimer’s disease and normal aging
- Wang, Rose, et al.
- 2000
(Show Context)
Citation Context ...t, the intensity variation among these voxels can largely be attributed to IIH. However, expert supervision to select points is time consuming and error prone, especially for volume data. Wang et al. =-=[59]-=- haspresentedanautomatedmethod for generating the reference points. Meyer et al. [49] employed the LCJ method [60] topreliminarily segment the image and then fit a smooth functional over the segmented... |