{"id":24105,"date":"2018-12-25T10:32:25","date_gmt":"2018-12-25T10:32:25","guid":{"rendered":"http:\/\/biomedpharmajournal.org\/?p=24105"},"modified":"2020-04-24T05:15:09","modified_gmt":"2020-04-24T05:15:09","slug":"probabilistic-identification-and-estimation-of-noise-application-to-mr-images","status":"publish","type":"post","link":"https:\/\/biomedpharmajournal.org\/staging\/vol11no4\/probabilistic-identification-and-estimation-of-noise-application-to-mr-images\/","title":{"rendered":"Probabilistic Identification and Estimation of Noise: Application to MR Images"},"content":{"rendered":"

Introduction<\/strong><\/p>\n

Noise assement in MRI 1,2,3<\/sup> usually requires the assessment of standard deviation of noise. Earlier methods can be separated into two methods for the computation of noise, first method involves the manually selected region of interest (ROI), in the second method entire volumetric data or image is considered for estimation without human interpretation. The problem is facing for the current automatic estimation method is separation of original signal from noise effected signals and other in homogeneity artifacts4,5<\/sup> the proposed work of two authors is by aggregate the values of all pixels from an entire pixel data set in to one dimensional array and estimate the standard deviation of noise from the histogram of one dimensional array.in this research work, introduce simpler method for noise assement in MR image to eliminate the drawback of two authors. The proposed method requires the recognition of noise only pixels which enhance the performance and accuracy of estimation of Gaussian noise and requires the prior knowledge about the standard deviation of Gaussian noise, and can cater distribution of noise for further analysis like segmentation and registration. Most of the noise estimation algorithms in the earlier method is based on background methods and object based methods, background methods which follow Rayleigh distributed noise and object based methods follow Gaussian and rician noise.6<\/sup> In this work we developed an algorithm for noise assement in MRI for the data acquired for multiple N receiver coils in which data follows nc-\u00a0distribution is the special case of Rician distribution.7<\/sup> Recently the author Rajan7,11<\/sup> developed an algorithm for the estimation of noise without background. Most of the above estimation algorithms are developed are sensitive to artifacts results in inaccuracy of noise estimation, the author coupe8<\/sup> developed robust noise estimation algorithm rician median absolute deviation [RMAD] for noise bias correction and artifacts to achieve accurate noise estimation, Other methods of noise estimation is variance stabilization transformation framework (VST) for bias inhomogeneity12,13<\/sup> correction in\u00a0 MR image to transform bias Rician noise distribution in to uniform Gaussian distribution and achieves accuracy noise estimation compared to other noise estimation methods.<\/p>\n

Materials and Method<\/strong><\/p>\n

The data set obtained from the DIPY (diffusion imaging in python) and JSS hospital Mysore Karnataka, India.<\/p>\n

Data set 1<\/strong><\/p>\n

STANFORD_HARDI (High resolution Diffusion Weighted imaging dataset (N=4)) Uses phased array coil system and SOS reconstruction (without parallel imaging).<\/p>\n

TAIWAN_NTU_DSI (Diffusion Spectrum imaging) Dataset (N=1) uses phased array coil system and SENSE reconstruction. (With parallel imaging).<\/p>\n

Dataset 2<\/strong><\/p>\n

Philips 3.0 Tscanner<\/strong><\/p>\n

MRI T1 weighted axial Brain image having pathology acquisition parameters are TR=5.3sec, TE=20ms, slice thickness=3.5mm, Resolution of 512×512. Parallel Image Reconstruction: SENSE.<\/p>\n

Philips 3.0Tscanner<\/strong><\/p>\n

MRI T1 weighted sagittal Brain image with acquisition parameters are TR=5.3 sec, TE=20ms, slice thickness=3.5mm, Resolution of 512×512. Parallel Image Reconstruction: SENSE.<\/p>\n

Diffusion Imaging in Python (Dipy) is a free and open source software project tool where image processing library tools are available for the analysis of data from diffusion magnetic resonance imaging (dMRI) experiments.<\/p>\n

Method<\/strong><\/p>\n

Noise modelling in MRI<\/strong><\/p>\n

The MR data acquired from the multiple N receiver coils m (or mi j k<\/sub>) follows non-central chi distribution [10] which is reconstructed from sum-of-squares algorithm (SOS) distribution,<\/p>\n

\"Vol11No4_Pro_Anj_for1\"<\/p>\n

distribution of 2N degrees of freedom with the non- centrality distribution parameter is given by<\/p>\n

\u03b72\u00a0<\/sup>\/\u03c3g<\/sub>2<\/sup>.\u00a0The probability density function (PDF) of<\/p>\n

\"Vol11No4_Pro_Anj_for2\"<\/p>\n

is given by [9]\n

\"Equation<\/p>\n

Where the PDF is zero for m < 0,\u00a0\u03b7<\/em>\u00a0is the underlying (combined) signal intensity,\u00a0\u03c3<\/em>g\u00a0<\/sub>is the standard deviation of Gaussian noise and Ik<\/sub> is the kth-order modified Bessel function of the first kind. We should note that magnitude MR signals reconstructed from other parallel image reconstruction techniques may not follow non-Central Chi distribution, Note that when N= 1, Eq. (1) reduces to the Rician PDF [2].<\/p>\n

In this research work, the PDF of magnitude noise (i.e.\u00a0\u03b7\u00a0<\/em>=0) is more relevant than the PDF of magnitude signal (\u03b7\u00a0\u2260\u00a0<\/em>0).\u00a0Therefore, we will derive the PDF of magnitude noise from Eq. (1) by setting the input signal to zero, i.e.\u00a0\u03b7\u00a0<\/em>=0,\u00a0then it shows that\u00a0 PDF of magnitude noise is given by<\/p>\n

\"Equation<\/p>\n

After IN-1 <\/sub>in Eq. (1) is replaced by first term Taylor expansion about input signal\u00a0\u03b7\u00a0<\/em>=0,<\/p>\n

Expressed by the following expression<\/p>\n

\"Equation<\/p>\n

When N=1, Eq. (2) simplifies to Rayleigh PDF.<\/p>\n

Then by a change of variables from m to<\/p>\n

\"Vol11No4_Pro_Anj_for3\"<\/p>\n

shows that variable t follows form of the particular Gamma PDF<\/p>\n

\"Equation<\/p>\n

The Gamma PDF, fy\u00a0<\/sub><\/em>as defined in [31]:<\/p>\n

\"Equation<\/p>\n

and\u00a0\u0393\u00a0is the Gamma function.<\/p>\n

Distribution of the Mean of Several Sampling Observed M-Values<\/strong><\/p>\n

The measurements m i, j, k<\/sub>\u2019s are defined in Eq. (2). We have chosen the arithmetic mean for the identification of noise-only pixels. If knows the distribution for the mean, which is depends on\u00a0\u03c3<\/em>g<\/sub>, we can decide for any observed mean sample came from the noise-only distribution. Which will provide the expression for the arithmetic mean of the several sampling observed values of m i, j, k <\/sub>s through the method of Characteristic function. In short, the derived random variable si, j<\/sub> representing the arithmetic mean of K independent Gamma Random variables {ti,j,1,<\/sub> ti,j,2,…. <\/sub>ti,j,k<\/sub>} is given by<\/p>\n

\"Equation<\/p>\n

Further, the new random variable si ,j <\/sub>related to K independent magnitude MR measurements, {mi,j,1,<\/sub> mi,j,2,<\/sub> mi,j,k<\/sub>} drawn from\u00a0 the distribution of magnitude noise is given by<\/p>\n

\"Equation<\/p>\n

Identification of the noise-only pixels using probabilistic method requires to define threshold on \u2018s\u2019\u00a0 i,e upper and lower threshold, so that a given proportion of all noise-only voxels or pixels fall between these two values. And we need to derive the cumulative distribution function (CDF), and it\u2019s inverse of the distribution of \u00a0\u2018s\u2019 to specify the lower and upper threshold values of s. Assume that N and K are known and the initial estimate of\u00a0\u03c3<\/em>g<\/sub>,\u00a0is required, is obtained by automatically using the simple search method are below.<\/p>\n

The probability distribution function and CDF of \u00a0\u2018s\u2019 is denoted by<\/p>\n

\"Vol11No4_Pro_Anj_for4\"<\/p>\n

the inverse of CDF \u2018s\u2019 are denoted symbolically by \u03bb= ps <\/sub>-1 <\/sup>(a | N, K).<\/em><\/p>\n

The CDF of \u2018s\u2019 and its inverse are given by<\/p>\n

\u03bb= ps <\/sub>-1 <\/sup>(ps <\/sub>( \u03bb| N, K).| N, K)\u00a0<\/em>and a\u00a0<\/em>= ps<\/sub>-1 <\/sup>(ps <\/sub>( a| N, K)| N, K)\u00a0<\/em>Note that\u00a0 Ps<\/sub>-1 <\/sup>(a|N,K)\u00a0<\/em>given by Inverse Gamma Regularized [NK,1-\u03b1]|K\u00a0\u00a0in Mathematica[14].The lower and upper threshold values of\u00a0\u2018s\u2019 are\u00a0\u03bb\u00a0–<\/sub>,\u00a0\u03bb\u00a0+\u00a0<\/sub>denoted \u00a0probabilities are\u00a0a<\/em>\/2 and 1-(a<\/em>\/2)\u00a0can be expressed in terms of the inverse CDF of \u2018s\u2019 given by\u00a0\u03bb\u00a0– <\/sub>=\u00a0Ps<\/sub>-1\u00a0<\/sup>(a\/2|N,K), <\/em>and\u00a0\u00a0\u03bb\u00a0+\u00a0<\/sub>Ps<\/sub>-1\u00a0<\/sup>(1-a\/2|N,K),\u00a0<\/em>Thus, a collection of K independent magnitude MR signals\u00a0{m i,j,1<\/sub>, m i,j,2<\/sub>, m i,j,k<\/sub><\/em>}\u00a0\u00a0from the acquisition \u00a0is decided to contain only noise if<\/p>\n

\"Vol11No4_Pro_Anj_for5\"<\/p>\n

which \u00a0satisfies the inequalities,\u00a0\u03bb\u00a0–\u00a0<\/sub>\u2264\u00a0S\u00a0i,j,\u00a0<\/sub><\/em>\u2264 \u03bb\u00a0+<\/sub>.<\/p>\n

Noise Estimation using the sample median, the sample mean or optimal quantiles method<\/strong><\/p>\n

The computation of Gaussian noise SD is obtained from the simple method of mean, denoted by <m>, of means of a collection of acquisition data measurements. The estimation of Gaussian noise SD,\u00a0\u03c3g<\/sub>,\u00a0can also be computed from the sample quantile of a unique optimal order \u03b1*<\/p>\n

\"Equation<\/p>\n

Note that both the sample median and the median of the continuous PDF, Eq. (2), are denoted by the same symbol. Note also that the denominator of Eq. (9) can be computed in advance, see Table 2. For example, when N = 1, Eq. (9) can be expressed mathematically<\/p>\n

\"Equation<\/p>\n

\"Vol11No4_Pro_Anj_for6\"<\/p>\n\n\n\n
\"Figure<\/td>\nFigure 1: Block diagram of PIESNO Algorithm.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

Algorithm Implementation<\/strong><\/p>\n

Input parameters:\u00a0\u00a0\u00a0 N, K,\u00a0\u03bb\u00a0–<\/sub>, \u03bb+<\/sub>,\u00a0<\/sub>m\u00a0i,j,k\u00a0\u00a0<\/sub><\/em>for all input measurements, the estimate of\u00a0\u03c3g<\/sub>,\u00a0is needed<\/p>\n

Output framework:\u00a0 identification of noisy only pixels and estimate of\u00a0\u03c3g<\/sub>,\u00a0and array of noise elements ‘\u03a9’<\/p>\n

Step 1 is the identification of noise-only pixels, and dynamically adding noisy pixels in to new elements of the one-dimensional array,\u00a0\u03a9.\u00a0And compute the standard deviation of noise in advance using the equation shown below.<\/p>\n

At each pixel location (i,j), compute si,j:<\/sub><\/p>\n

\"Equation<\/p>\n

A positive identification is made if\u00a0S\u00a0i,j<\/sub><\/em> satisfies the condition\u00a0\u00a0\u03bb\u00a0–\u00a0<\/sub>\u2264\u00a0S\u00a0i,j,\u00a0<\/sub><\/em>\u2264 \u03bb\u00a0+<\/sub>.<\/p>\n

In Step 2, the channel between the identification and estimation of noise-only pixels is set and create one-dimensional array of\u00a0\u03a9,\u00a0which is the union of all the arrays of K measurements identified as noise only pixels. Note that the number of positive identifications is defined is equal to the number of the positively identified arrays. Finally, the sample median of\u00a0\u03a9, 1,\u00a0is selected. Note again that the sample mean of\u00a0\u03a9\u00a0or the sample quantile of\u00a0\u03a9\u00a0of a specific order may be used here.<\/p>\n

Step 3 is the computation of standard deviation of noise, the array of sample median, l, of\u00a0\u00a0\u03a9\u00a0computed in Step 2 is used to estimate\u00a0\u03c3g.\u00a0<\/sub>Note again that the sample mean and other optimal quantiles are defined below.<\/p>\n

\"Equation<\/p>\n

In Step 4, if the iteration stops the \u03a9 is empty, which repeats procedure from Steps 1 to 4 sequentially until convergence or the maximum number of iterations is reached. Here, a sequence of estimates of\u00a0\u03c3g\u00a0<\/sub>produced by the iterative procedure is considered if the absolute difference between any successive estimates is less than small number, say 0.00000000001.<\/p>\n

Step 5, the noise identification and computation of the standard deviation of noise\u00a0 is denoised using NLM algorithm and evaluate its performance parameters.<\/p>\n\n\n\n
\"Table<\/td>\nTable 1: The optimal quantile value for different N and the value of the Denominator of Eq. (8) for different N but at the optimal quantile order.<\/strong><\/p>\n

 <\/p>\n

Click here to view\u00a0table<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\u00a0\"Table<\/td>\nTable 2: The denominator of Eq. (9) values for various N.<\/strong><\/p>\n

 <\/p>\n

Click here to view\u00a0table<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

Results and Discussion<\/strong><\/p>\n

The proposed algorithm is implemented using Python. After the python implementation of the PIE S. NO we plot the SD of noise-only pixels using MATPLOTLIB with N (phased array coils) varying from 1 to 12. The figure (4-6) below shows the plots of PDF of noise-only pixels vs SD (sigma) with N=1, 4 and 10. Using different data sets below.<\/p>\n

Piesno Results for Different Datasets<\/strong><\/p>\n

Tested on Taiwan and Stand ford Dataset<\/strong><\/p>\n

Fig.2 and fig.3 shows the data set obtained from Stanford_hardi dataset acquired using 4 phased array coils. In these images the White dots appearing in the image are the noise-only pixels identified and estimated using our proposed PIESNO algorithm. These noisy only pixels are denoised using well known NLM algorithm using python. The image is subjected to denoising with Background preservation. Similarly the data set obtained from Taiwan dataset acquired using single phase array coil, from these two graphs we observed that the noise identified is Rician and Gaussian from the background region and it is well modelled by Gaussian and rician from the PDF curve is as shown in fig.4 and fig.5.<\/p>\n\n\n\n
\"Figure<\/td>\nFigure 2: Stanford_Hardi Dataset (N=4).<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\"Figure<\/td>\nFigure 3: Taiwan_Ntu_Dsi Dataset (N=1).<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\"Figure<\/td>\nFigure 4: N=1 Rician distribution.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

From the algorithm the estimated value of standard deviation of noise and its PDF curve is shown in fig.4 above. The curve is not symmetric about a mean value. Hence we identify the noise present in the MRI as Rician Noise or nc-\u03bb distribution.<\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 5: N=4 Gaussian distribution.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

For the fig.4 above estimated value of standard deviation of noise using PIESNO algorithm using N=4, the plot obtained is a symmetric curve as shown above. The curve is symmetric about the value 0. Hence it follows a normal distribution and the noise is identified as a Gaussian Noise.<\/p>\n

Similarly for N>1, the plot follows the same curve as shown below fig.5 for N=10.<\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 6: N=10 Gaussian distribution.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

Thus we can interpret the results of PIESNO as<\/p>\n

For N=1 the noise estimated and identified is Rician Noise.<\/p>\n

For N>1 the noise estimated and identified is Gaussian Noise.<\/p>\n

Tested on Clinical Data<\/strong><\/p>\n

Fig.7\u00a0 shows the data acquired from MR machine (clinical data) using multiple phase array coils using parallel image reconstruction SENSE algorithm and fig. 8\u00a0 shows data acquired using SOS reconstruction algorithm without subsampling. The identification of noise distribution using PIESONO algorithm and its results with histograms is as shown in fig. 9-12, from this graphical analysis found that the noise distribution is Gaussian and rician.<\/p>\n

The denoising results of the algorithm for Taiwan and stand ford dataset are shown in fig.13 and fig.14 and difference image (Residual image) is shown, it is observed that the algorithm shows better preservation of edge and efficient noise reduction.<\/p>\n

The comparison of denoising algorithm with other algorithm is shown in table 3. From this table it is observed that the NLM algorithm shows efficient noise reduction with others and it is verified from PSNR values, and it is burden due to long execution time compared with AONLM and PCA algorithm is shown in the table 3.<\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 7: MRI T1 Weighted sequence images in the axial and sagittal view acquired using multiple Phase Array coils (parallel image reconstruction: SENSE).<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 8:<\/strong> MRI T1 Weighted sequence images in the axial view acquired using multiple phase Array coils (N=4, SOS reconstruction without subsampling).<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\"Figure<\/td>\nFigure 9: Histogram of fig.7 (a) shows the distribution is approximated to Rician distribution.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 10: Histogram of fig.7 (b) shows the distribution is approximated to Rician distribution.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\"Figure<\/td>\nFigure 11: Histogram of fig.8 (a) shows the distribution is approximated to Gaussian distribution.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 12:\u00a0<\/strong>Histogram of fig. 8 (b) shows the distribution is approximated to Gaussian distribution.<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 13: Denoised Result for Stanford_Hardi (N=4).<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\u00a0\"Figure<\/td>\nFigure 14: Denoised Result of Taiwan_Ntu_Dsi (N=4).<\/strong><\/p>\n

 <\/p>\n

Click here to view figure<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n\n\n\n
\"Table<\/td>\nTable 3: The comparison of denoising methods with different noise levels on MR T1 weighted Images.<\/strong><\/p>\n

 <\/p>\n

Click here to view\u00a0table<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

Discussion<\/strong><\/p>\n

Earlier estimation algorithms work for single image but our proposed algorithm is well suited for multiple images to estimate the variance of noise and it is not suited for clinical practice requires longer acquisition time to estimate the level of noise.<\/strong> Our method is applicable to larger class of magnitude data measurements acquired from the MR machine than Rayleigh-distributed data (i.e., N = 1) and discuss some limitations of the proposed approach. PIESNO cannot be expected to perform well for very small K because the identification of noise may be less stable and may not produce sufficient number of elements for the estimation of noise via the median method to be useful as an estimator. We tested the effectiveness of the algorithm with clinical and synthetic database and conclude that stastical distribution of noise is Gaussian, rician or nc-\u03bb distribution and it is verified from mathematical derived probabilistic distribution function. The denoising performance of the algorithm is compared with standard conventional algorithm and execution time.<\/p>\n

References<\/strong><\/p>\n

    \n
  1. Macovski. Noise in MRI, Magn. Reson. Med.<\/em> 1996;36:494\u2013497.
    \n    CrossRef<\/a><\/li>\n
  2. Gudbjartsson H.,Patz\u00a0S. The Rician distribution of noisy MRI data.Magnetic Resonance in Medicine<\/em>. 1995;34(6)910\u2013914.
    \n    CrossRef<\/a><\/li>\n
  3. Drumheller. General expressions for Rician density and distribution functions.\u00a0IEEE Trans. on Aerospace and Electronic Systems.<\/em> 1993;29(2):580\u2013588.
    \n    CrossRef<\/a><\/li>\n
  4. Sijbers.,\u00a0 den Dekker J. A., Van Verhoye A. M.,Van Dyck D. Estimation of the noise in magnitude MR images. Magn. Reson.Imag<\/em>. 1998;16(1):87\u201390.<\/li>\n
  5. Sijbers., Poot D., den Dekker A. J and Pintjens W.\u00a0 Automatic estimation of the noise variance from the histogram of a magnetic resonance image. Physics in Medicine and Biology.<\/em> 2007;52(5):1335-1348.
    \n    CrossRef<\/a><\/li>\n
  6. Aja-Fernandez., Niethammer M., Kubicki M., Shenton M. E and Westin C. F. Restoration of dwi data using a rician lmmse estimator. Medical Imaging. IEEE Transactions on.<\/em> 2008;27(10):1389-1403.<\/li>\n
  7. Aja-Fernandez.,\u00a0 Alberola-Lopez C and Westin\u00a0C. F. Noise and signal estimation in magnitude MRI and Rician distributed images: a LMMSE approach. IEEE transactions on imag<\/em>e processing.\u00a0 2008;17(8):1383-1398.
    \n    CrossRef<\/a><\/li>\n
  8. Sijbers.,\u00a0 den Dekker A.,\u00a0 Scheunders P.,Van Dyck D. Maximum-likelihood estimation of Rician distribution parameters. IEEE Trans. Med.Imaging.<\/em> 1998;17(3):357\u2013361.
    \n    CrossRef<\/a><\/li>\n
  9. \u00a0Koay G., Basser P. J. Analytically exact correction scheme for signal extraction from noisy magnitude MR signals.Journal of Magnetic Resonance<\/em>. 2006;179(2):317\u2013322.
    \n    CrossRef<\/a><\/li>\n
  10. Aja-Fernandez., Tristan-Vega A.,\u00a0 Alberola-Lopez C. Noise estimation in single- and multiple-coil magnetic resonance data based on \u00b4statistical models. Magn. Reson. Imag.<\/em> 200;27:1397\u20131409.<\/li>\n
  11. Rajan.,\u00a0 Poot D., Juntu J.,\u00a0 Sijbers J. Noise measurement from magnitude MRI using local estimates of variance and skewness. Physics in Medicine and Biology.<\/em> 2010;55(16):441.
    \n    CrossRef<\/a><\/li>\n
  12. Coup\u00e9 P.,\u00a0 Manj\u00f3n J., Gedamu E.,\u00a0 Arnold D., Robles M and Collins D. L. Robust Rician noise estimation for MR images. Medical Image Analysis.<\/em> 2010;14(4):483\u2013493.
    \n    CrossRef<\/a><\/li>\n
  13. Foi A. Noise estimation and removal in MR imaging: the variance-stabilization approach,” in [Proceedings of the IEEE International Symposium on Biomedical Imaging: From Nano to Macro]. 2011.
    \n    CrossRef<\/a><\/li>\n
  14. \u00a0Wolfram S. The Mathematica Book, Wolfram Media, Champaign. 2003.<\/li>\n<\/ol>\n","protected":false},"excerpt":{"rendered":"

    Introduction Noise assement in MRI 1,2,3 usually requires the assessment […]<\/p>\n","protected":false},"author":9,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[61],"tags":[],"class_list":["post-24105","post","type-post","status-publish","format-standard","hentry","category-vol11no4"],"_links":{"self":[{"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/posts\/24105","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/users\/9"}],"replies":[{"embeddable":true,"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/comments?post=24105"}],"version-history":[{"count":6,"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/posts\/24105\/revisions"}],"predecessor-version":[{"id":32590,"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/posts\/24105\/revisions\/32590"}],"wp:attachment":[{"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/media?parent=24105"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/categories?post=24105"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/biomedpharmajournal.org\/staging\/wp-json\/wp\/v2\/tags?post=24105"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}