CN103793711A - Multidimensional vein extracting method based on brain nuclear magnetic resonance image - Google Patents

Abstract

The invention discloses a multidimensional vein extracting method based on a brain nuclear magnetic resonance image. An area-of-interests in the brain nuclear magnetic resonance image is segmented through a region growing method. Vein characteristic parameters of the area-of-interests are extracted through a Curvelet conversion and Contourlet conversion method. The people comprise an Alzheimer patient group, a mild cognitive impairment patient group and normal old people group, the vein characteristic parameters of the area-of-interests comprise entropy, gray average, correlation, energy, the homogeneity degree, variance, the maximum probability, deficit distance, an inverse gap, clustering tendency, contrast, a sum mean value, a difference mean value, a sum entropy and a difference entropy. The area-of-interests comprises a entorhinal cortex are and a sea horse area.

Description

A kind of various dimensions texture extracting method based on brain nuclear magnetic resonance image

Technical field:

The invention belongs to medicine technology field, be specifically related to a kind of various dimensions texture extracting method based on brain nuclear magnetic resonance image (MRI).

Background technology:

In the early stage Alzheimer disease of auxiliary diagnosis (AD), in identification MRI image, ROIs(comprises entorhinal cortex, hippocampus) character significant.But MRI image technology can only be using atrophy of hippocampal as distinguishing one of index of patient and normal person, and doctor is subject to subjective personal influence to the explanation of MRI image, is lack of consistency, and is difficult for the order of severity of accurate evaluation dementia patients symptom.

1, existing image processing techniques has following 5 kinds:

1) region growing method (Region-growing Method):

The method has been utilized the local spatial information of image, can effectively overcome the discontinuous shortcoming of image partition space that other method exists, but the method that is still out of use is carried out the processing of brain MRI image.

2) gray level co-occurrence matrixes (GLCM):

Only had the textural characteristics parameter that adopts gray level co-occurrence matrixes method to extract in a small amount of correlative study, this is for diagnosing early stage AD, MCI to be also nowhere near according to brain MRI image texture characteristic in the past.

3) wavelet transformation (Wavelet Transformation):

Although the proper vector that wavelet transformation forms can accurately be described image to a certain extent, utilize wavelet transformation to extract ROIs textural characteristics in image and exist the shortcoming that retrieval precision is not high.

4) Second Generation Wavelet Transformation (Curvelet conversion):

After the wavelet transformation growing up continue the later stage eighties in last century, within 1996, Swendens has proposed advanced Second Generation Wavelet Transformation, on basis function algorithm, also updating, E.J.Candes in 1998 has proposed Ridgelet conversion, E.J.Candes in 1999 and D.L.Donoho and has invented Curvelet conversion new algorithm:

Wherein: 2 ^-jfor yardstick, θ ^lfor deflection θ ^l,

for position

r is conversion radian, within 2006, has proposed again fast discrete Curvelet conversion.Second Generation Wavelet Transformation has not only retained the multiple dimensioned advantage of wavelet transformation (Wavelet Transformation) method, also there is anisotropy feature simultaneously, can approach well singular curve, more be applicable to curve or the edge feature in analysis of two-dimensional images than Wavelet Transformation, and there is higher approximation accuracy and better sparse expression ability, can provide a kind of than the more accurate method of Wavelet Transformation multiresolution analysis for image.

5) Contourlet conversion

The anisotropy scaling relation of Curvelet conversion has been inherited in Contourlet conversion, and it is the another kind of implementation of Curvelet conversion in a sense.The basic thought of Contourlet conversion is that first the multiple dimensioned decomposition with a similar small echo catches edge singular point, then according to directional information, singular point close position is accumulated to contour segment.

Contourlet conversion can be divided into two parts: the tower filter construction of Laplce (Laplacian Pyramid, LP) and two-dimensional directional bank of filters (Directional Filter Bank, DFB).LP decompose that low pass sampling that first produces original signal approaches and original image and low pass predicted picture between an error image, the low-pass pictures obtaining is continued to decompose the low-pass pictures and the error image that obtain lower one deck, and so progressively filtering obtains the Multiresolution Decomposition of image; DFB bank of filters is used the conjugate mirror filter group of sector structure to avoid the modulation to input signal, the anisotropic filter of 1 layer of y-bend tree structure has been become to the structure of 21 parallel channels simultaneously.

Contourlet conversion is a kind of new image two-dimensional representation method, there is the character such as multiresolution, local positioning, multidirectional, neighbour's circle sampling and anisotropy, its basis function be distributed in multiple dimensioned, multi-direction on, a small amount of coefficient can catch the edge contour in image effectively, and the edge contour principal character in image just.

But these new methods, in the time processing the MRI image of different parts, need to be utilized basis function to re-construct new algorithm, choose suitable parameter, therefore still have many theoretical questions to be worth research.Contourlet conversion has been successfully used to the practical problemss such as image co-registration, and for the bibliographical information phoenix feathers and unicorn horns of brain image texture feature extraction.In the document of consulting at present, only someone uses GLCM and Wavelet transfer pair AD group and normal group brain MRI image texture feature extraction and sets up forecast model, with diagnostic result accuracy, above-mentioned two kinds of methods are compared, find to adopt GLCM to extract texture modeling and forecasting effect and be better than Wavelet conversion.Having not yet to see someone uses Second Generation Wavelet Transformation and Contourlet conversion to carry out the extraction of AD brain MRI image texture.Therefore, need to be in conjunction with the advantage of above-mentioned prior art, and overcome its deficiency, the disposal route of image is improved, to reach the object that improves early stage AD, MCI diagnosis.

2, conventional forecast model:

Support vector machine (Support Vector Machine, SVM):

Support vector machine is the machine learning method being based upon on Statistical Learning Theory VC dimension theory and structural risk minimization principle basis.Its mechanism is to find an optimal classification lineoid that meets classificating requirement, makes this lineoid in guaranteeing nicety of grading, can make the white space of lineoid both sides maximize.

In theory, support vector machine can realize the optimal classification to linear separability data.Take two class Data classifications as example, given training sample set (x _i, y _j), i=1,2 ..., l, { ± 1}, lineoid is denoted as (wx)+b=0 to x ∈, for making classifying face correctly classify to all samples and possess class interval, just requires it to meet following constraint: y _i[(wx _i)+b]>=1, i=1,2 ..., l, can calculate class interval is 2/||w||, the problem of therefore constructing optimum lineoid is just converted under constraint formula to be asked:

$\min \mathrm{\φ}\left(w\right)=\frac{1}{2}{\left|\right|w\left|\right|}^2=\frac{1}{2}(w^{,}\·w)$

In order to solve this constrained optimization problem, introduce Lagrange function:

$L(w,b,a)=\frac{1}{2}\left|\right|w\left|\right|-\mathrm{\α}(y((w\·x)+b)-1)$

In formula, α _i> 0 is Lagrange multiplier.The solution of constrained optimization problem determines by the saddle point of Lagrange function, and the solution of optimization problem to meet at saddle point place be 0 to the local derviation of w and b, this QP problem is converted into corresponding dual problem:

$\max Q\left(\mathrm{\α}\right)=\underset{j=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\α}}_j-\frac{1}{2}\underset{i=1}{\overset{l}{\mathrm{\Σ}}}\underset{j=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_j(x_i\·x_{\text{j}})$

$s.t.\underset{j=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\α}}_j{y}_j=0,j=\mathrm{1,2},...,l$

Solve optimum solution ${\mathrm{\α}}^{*}={({\mathrm{\α}}_1^{*},{\mathrm{\α}}_2^{*},...,{\mathrm{\α}}_l^{*})}^l$

Calculate optimum weight vector w ^*with optimum biasing b ^*, be respectively:

$w^{*}=\underset{j=1}{\overset{1}{\mathrm{\Σ}}}{\mathrm{\α}}_j^{*}{y}_j{x}_j$

$b^{*}=y_i-\underset{j=1}{\overset{l}{\mathrm{\Σ}}}{y}_j{\mathrm{\α}}_j^{*}(x_j\·x_i)$

In formula, subscript

therefore obtain optimal classification lineoid

(w ^*x)+b ^*=0, and optimal classification function is:

$f\left(x\right)=\mathrm{sgn}\{(w^{*}\·x)+b^{*}\}=\mathrm{sgn}\{\underset{j=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\α}}_j^{*}{y}_j(x_j\·x_i)+b^{*}\},x\∈R^n$

For linearly inseparable situation, the main thought of SVM is by the characteristic vector space of defeated people's DUAL PROBLEMS OF VECTOR MAPPING to higher-dimension, and in this feature space, constructs optimal classification face.

X is done from input space R ⁿto the conversion φ of feature space H:

x→φ(x)＝（φ ₁（x）,φ ₂(x),…φ _l(x)） ^T

Replace input vector x with proper vector φ (x), can obtain optimal classification function and be:

$f\left(x\right)=\mathrm{sgn}(w\·\mathrm{\φ}\left(x\right)+b)=\mathrm{sgn}(\underset{j=1}{\overset{l}{\mathrm{\Σ}}}{a}_j{y}_i\mathrm{\φ}\left(x_i\right)\·\mathrm{\φ}\left(x\right)+b)$

Summary of the invention:

The object of the present invention is to provide a kind of to containing the dividing method of MRI image and the method for texture feature extraction of MCI, early stage AD patient's focus and Elderly people crowd ROIs, set up multiple forecast model, so that more effective discovery MCI patient, diagnoses early stage AD and observation Elderly people crowd's all brain structures to change.

The present invention combines the advantage of the methods such as region growing method of the prior art, gray level co-occurrence matrixes, wavelet transformation, and improved, use Curvelet conversion and Contourlet conversion to extract the Edge texture feature of ROIs, texture extracting method is comprehensive, novel, can reach goal of the invention.

The present invention has set up a kind of to the cutting apart and the method (specific procedure is shown in Fig. 4) of texture feature extraction of the MRI image that contains MCI, early stage AD patient's focus and normal aging people ROIs by following main technological route, and has set up accordingly the forecast model of the relevant ROIs character of judgement:

1) set up ROIs image library;

2) deployment area growth method splits relevant ROIs in image;

3) adopt Curvelet conversion and Contourlet conversion process image, extract following variable: entropy (entropy), gray average (gray average), correlativity (correlativity), energy (energy), homogeneity degree (homogeneity degree geneity), variance (Variance), maximum probability (Maximum probability, maximum probability), unfavourable balance is apart from (Inverse Difference Moment, unfavourable balance distance), Clustering Tendency (Cluster Tendency), contrast (Contrast), with average (Sun-gray average), poor average (Difference-gray average), with entropy (Sum-entropy), poor entropy (Difference-entropy),

4) by step 2)～3) the various variable data that obtain set up characteristics of image parametric data storehouse;

5) according to step 4) database build the forecast model converting based on Curvelet conversion and Contourlet; The method of setting up forecast model comprises support vector machine;

6) by step 5) the various parametric data that obtain and sample be through checking repeatedly, to revise forecast model, obtains result ideal model more accurately.

7) comparison step 5) in convert based on Curvelet conversion and Contourlet the prediction effect of setting up forecast model.

At present for the also unified definition of neither one of texture of image, what it is generally acknowledged that the textural characteristics of image describes is the variation of body surface gray scale or color, and this variation is relevant with the self attributes of object, is the repetition of certain texture primitive.

Adopt Curvelet conversion and Contourlet conversion can obtain following textural characteristics parameter:

Entropy entropy:

the randomness of reflection image texture;

Gray average gray average:

the central tendency of all gray-scale values of reflection pixel.

Correlativity correlativity: measure the correlativity of pixel grey scale;

Energy (Angular Second Moment) energy (angle second moment):

degree of uniformity and texture fineness degree that reflection image greyscale distributes;

Homogeneity degree geneity homogeneity degree:

reflection grey level similarity degree;

Variance variance:

reflection grey level's distribution situation;

Maximum probability (maximum probability) maximum probability:

the incidence that the most outstanding pixel is right;

Inverse Difference Moment (unfavourable balance distance) unfavourable balance square:

the flatness of reflection image;

Cluster tendency Clustering Tendency:

measure similar grey level and be worth the grouping of pixel;

Contrast contrast:

the sharpness of reflection image;

Sun-gray average and average

the average that Difference-gray average is poor

the average of grey level in image is provided.

Sum-entropy and entropy

the entropy that Difference-entropy is poor $f_9=-\underset{k=0}{\overset{K-1}{\mathrm{\Σ}}}{c}_{x-y}\left(k\right)\log \left\{c_{x-y}\left(k\right)\right\}.$

The present invention has set up early stage AD forecast model with said method, and accuracy of judgement degree has reached 100%.

Be below the example that ROIs inner vein extracts, step is as follows:

1, collect the MRI original image (accompanying drawing explanation is take the MRI image of normal aging people as example) of AD, MCI and normal aging people, see Fig. 1;

2, cut apart above-mentioned image by region growing method, obtain image and see Fig. 2, Fig. 3, cut apart the program writing that adopts, directly operation.

3, adopt Curvelet conversion and Contourlet conversion texture feature extraction parameter, every kind of method has respectively corresponding program, directly operation, and the textural characteristics parameter of extraction is in table 2～table 29.

The results show:

Table 1Curvelet conversion changes different azimuth difference analysis with Contourlet

As shown in Table 1, Contourlet ratio of transformation Curvelet conversion can better reflect AD group, MCI group, texture value total difference between normal group.

Adopt region growing method to cut apart ROIs, choose 80% data as training sample, again according to remaining 20% data as checking sample, judge the whether consistance of pathology and pathological diagnosis of ROIs with the model of upchecking, prove that extracting brain MRI image ROIs external texture by Contourlet sets up sensitivity, the specificity of the prediction early stage of lung cancer and be 100%.

By above data, can obtain conclusion: adopt region growing method to cut apart brain MRI image, and logical Contourlet conversion texture feature extraction parameter is set up forecast model early stage AD auxiliary diagnosis is had to good effect.

Beneficial effect:

Region growing method is used for to cutting apart of brain MRI, and this is an innovation of the present invention.The experiment proved that, region growing method is cut apart and is extracted texture modeling and will more be better than entirety and cut apart and extract texture modeling, can better retain the marginal information of tubercle;

Contourlet is the effective ways that extract brain MRI image inner vein feature, can extract 14 kinds of textural characteristics parameters, more comprehensively reflects the textural characteristics of image.

Accompanying drawing explanation:

Fig. 1 is the MRI original image of normal aging people, and that wherein mark with black surround is ROIs;

Fig. 2 is the left brain ROIs enlarged image obtaining after cutting apart by region growing method in Fig. 1;

Fig. 3 is the right brain ROIs enlarged image obtaining after cutting apart by region growing method in Fig. 1;

Fig. 4 is the technology path that brain MRI image is carried out to the extraction of various dimensions texture by the inventive method;

Wherein, the 1st, left cerebral hippocampal, entorhinal cortex region; The 2nd, right cerebral hippocampal, entorhinal cortex region.

Embodiment

Following instance is the model introduction of setting up the brain MRI image prediction AD based on containing relevant ROIs by the inventive method, and this is just to the further illustrating of the inventive method, but example does not limit range of application of the present invention.In fact, also can carry out character to the medical image of other type by this method judges.

Image source: shared AD, MCI and the brain MRI image of normal aging people on ADNI website, be respectively the image of .Nii form, use MRIcro software to read;

Method: adopt Matlab software programming, deployment area growth method splits the ROIs in above-mentioned MRI image, utilizes Contourlet conversion to extract the textural characteristics parameter of relevant ROIs.

Be below the example that brain MRI image ROIs textural characteristics parameter extracts, step is as follows:

1, collect respectively AD group case 20 examples, MCI case group 20 examples, normal group case 20 examples 250 brain MRI original images altogether, the case age is 40～89 years old, and the mean age is 64 years old, and the median of age distribution is 60 years old.

2, cut apart image by region growing method, partitioning scheme is pressed region growing method conventional method, adopts the program writing, and setting threshold is 35, directly operation.Fig. 1 is that an image to be split is given an example, obtains image and sees Fig. 2, Fig. 3,

3, adopt respectively Curvelet conversion and Contourlet conversion texture feature extraction parameter, the textural characteristics parameter of extraction is in table 2～table 29.

4, adopt Curvelet conversion and Contourlet transform method extraction brain MRI image ROIs to obtain respectively 14 textural characteristics parameters and set up forecast model.

5, comparison Curvelet conversion and two kinds of texture extracting method of Contourlet conversion are set up the prediction effect of forecast model.

By 200 brain MRI images of 12 example, as training sample set, remaining 50 brain MRI images of 3 example carry out the support vector machine prediction of classifying, prediction effect: sensitivity=100% as test sample book; Specificity=100%; Coincidence rate=100%.

The results show: adopt region growing method to cut apart ROIs, choose 80% data as training sample, again according to 20% of remainder data as checking sample, judge the whether consistance of pathology and pathological diagnosis of lung ROIs with the model of upchecking, prove by small sample brain MRI image ROIs texture feature extraction parameter is predicted to the sensitivity of early stage AD is 100%.

By above data, can obtain conclusion: adopt region growing method to cut apart brain MRI image ROIs, and convert texture feature extraction parameter by Contourlet and set up forecast model the auxiliary diagnosis of early stage AD is had to good effect.

The variable that Curvelet conversion is extracted is in table 2～table 15, and in table 16～table 29(, the case texture value take AD group # as 136_s_0194 is example to the variable that Contourlet conversion is extracted).

Table 2 gray average

Table 3 standard deviation

Standard deviation 1	Standard deviation 2	Standard deviation 3	Standard deviation 4	Standard deviation 5	Standard deviation 6
						4.3847711	1.42498491	1.36818905	1.41228397	1.462426799	1.591283749
Standard deviation 7	Standard deviation 8	Standard deviation 9	Standard deviation 10	Standard deviation 11	Standard deviation 12
						1.32885786	1.9148351	1.41065529	1.55030154	1.367888924	1.405432669
Standard deviation 13	Standard deviation 14	Standard deviation 15	Standard deviation 16	Standard deviation 17	Standard deviation 18
						1.30044057	1.579707287	1.33576328	1.18453501	1.427709383	1.277003212
Standard deviation 19	Standard deviation 20	Standard deviation 21	Standard deviation 22	Standard deviation 23	Standard deviation 24
						1.7011195	1.443854262	1.38209543	1.57756675	1.392241419	1.647125047
Standard deviation 25	Standard deviation 26	Standard deviation 27	Standard deviation 28	Standard deviation 29	Standard deviation 30
						1.42259127	1.430469425	1.51028701	1.27215853	1.311299465	1.527251474
Standard deviation 31	Standard deviation 32	Standard deviation 33	Standard deviation 34	?	?
						1.04393555	1.52005899	1.7647135	0.86030865	?	?

Table 4 Clustering Tendency

Clustering Tendency 1	Clustering Tendency 2	Clustering Tendency 3	Clustering Tendency 4	Clustering Tendency 5	Clustering Tendency 6
						16.3606301	16.12439285	18.3695153	15.9238466	13.85499042	21.2569512
Clustering Tendency 7	Clustering Tendency 8	Clustering Tendency 9	Clustering Tendency 10	Clustering Tendency 11	Clustering Tendency 12
						18.661451	18.32069429	16.9243141	17.3187788	14.44784022	20.6990795
Clustering Tendency 13	Clustering Tendency 14	Clustering Tendency 15	Clustering Tendency 16	Clustering Tendency 17	Clustering Tendency 18

11.959189	14.13415804	17.3955427	17.5325772	16.98576286	14.217201
						Clustering Tendency 19	Clustering Tendency 20	Clustering Tendency 21	Clustering Tendency 22	Clustering Tendency 23	Clustering Tendency 24
16.8114409	15.24536246	15.3914861	15.5040628	15.43258971	13.98916839
						Clustering Tendency 25	Clustering Tendency 26	Clustering Tendency 27	Clustering Tendency 28	Clustering Tendency 29	Clustering Tendency 30
16.331876	19.96832162	14.7059386	16.0655662	16.21379157	14.04231151
						Clustering Tendency 31	Clustering Tendency 32	Clustering Tendency 33	Clustering Tendency 34	?	?
18.2942587	16.86866386	16.5871928	16.2555715	?	?

Table 5 homogeneity degree

Table 6 maximum probability

Table 7 energy

Energy 1	Energy 2	Energy 3	Energy 4	Energy 5	Energy 6
						0.0498905	0.1160032	0.194533	0.173957	0.2182537	0.12101859
Energy 7	Energy 8	Energy 9	Energy 10	Energy 11	Energy 12
						0.1489191	0.0817874	0.091902	0.1012048	0.1858977	0.10748128
Energy 13	Energy 14	Energy 15	Energy 16	Energy 17	Energy 18
						0.345414	0.2464471	0.154017	0.1991889	0.1164763	0.12492103

Energy 19	Energy 20	Energy 21	Energy 22	Energy 23	Energy 24
						0.0749974	0.1107258	0.164832	0.1727838	0.1557092	0.14539656
Energy 25	Energy 26	Energy 27	Energy 28	Energy 29	Energy 30
						0.1146461	0.1371698	0.110386	0.1671202	0.4645303	0.29653236
Energy 31	Energy 32	Energy 33	Energy 34	?	?
						0.3057156	0.0939511	0.084532	0.5190262	?	?

Table 8 moment of inertia

Table 9 unfavourable balance distance

Table 10 entropy

Entropy 1	Entropy 2	Entropy 3	Entropy 4	Entropy 5	Entropy 6
						1.22035151	1.029386659	0.82359602	0.99692385	0.874746461	1.07754384
Entropy 7	Entropy 8	Entropy 9	Entropy 10	Entropy 11	Entropy 12
						0.93065033	1.278111547	1.12947679	1.12073057	0.887906447	0.932229361
Entropy 13	Entropy 14	Entropy 15	Entropy 16	Entropy 17	Entropy 18
						0.82547592	0.948837401	0.85367573	0.83820976	0.894950428	1.001405253
Entropy 19	Entropy 20	Entropy 21	Entropy 22	Entropy 23	Entropy 24
						1.09676062	1.008200647	0.81700194	0.90675533	0.978592281	1.211409108
Entropy 25	Entropy 26	Entropy 27	Entropy 28	Entropy 29	Entropy 30
						1.1444441	0.98236792	1.07274302	0.89660664	0.630961017	0.800516014

Entropy 31	Entropy 32	Entropy 33	Entropy 34	?	?
						0.58298469	1.029947319	1.1837409	0.5506891	?	?

Table 11 correlativity

Correlativity 1	Correlativity 2	Correlativity 3	Correlativity 4	Correlativity 5	Correlativity 6
						0.9155325	0.0907211	0.025449	-0.199867	-0.036875	-0.0596785
Correlativity 7	Correlativity 8	Correlativity 9	Correlativity 10	Correlativity 11	Correlativity 12
						-0.201381	-0.014917	-0.0417	0.0605341	0.0357446	-0.1837574
Correlativity 13	Correlativity 14	Correlativity 15	Correlativity 16	Correlativity 17	Correlativity 18
						0.0062253	-0.031564	-0.19318	0.0178852	0.0717881	0.02910943
Correlativity 19	Correlativity 20	Correlativity 21	Correlativity 22	Correlativity 23	Correlativity 24
						0.0133719	-0.211466	0.019258	0.016919	-0.191599	-0.00425
Correlativity 25	Correlativity 26	Correlativity 27	Correlativity 28	Correlativity 29	Correlativity 30
						-0.035369	0.0656479	0.010318	-0.184086	0.0910757	0.04394031
Correlativity 31	Correlativity 32	Correlativity 33	Correlativity 34	?	?
						-0.222713	-0.009445	0.084592	-0.006333	?	?

Table 12 and mean

The mean that table 13 is poor

Table 14 and entropy

The entropy that table 15 is poor

Poor entropy 1	Poor entropy 2	Poor entropy 3	Poor entropy 4	Poor entropy 5	Poor entropy 6
						2.7717822	2.7334586	2.646782	2.852217	2.7806391	2.90563906
Poor entropy 7	Poor entropy 8	Poor entropy 9	Poor entropy 10	Poor entropy 11	Poor entropy 12
						2.852217	3.25	2.646782	2.977217	3	2.77439747
Poor entropy 13	Poor entropy 14	Poor entropy 15	Poor entropy 16	Poor entropy 17	Poor entropy 18
						2.375	2.7806391	2.774397	2.4746018	2.9056391	2.64678222
Poor entropy 19	Poor entropy 20	Poor entropy 21	Poor entropy 22	Poor entropy 23	Poor entropy 24

3.0778195	2.6467822	2.608459	2.6467822	2.7743975	2.89939747
						Poor entropy 25	Poor entropy 26	Poor entropy 27	Poor entropy 28	Poor entropy 29	Poor entropy 30
2.6467822	2.7806391	2.774397	2.7743975	2.4362781	2.64939747
						Poor entropy 31	Poor entropy 32	Poor entropy 33	Poor entropy 34	?	?
2.5306391	2.7334586	3.024397	2.2169172	?	?

Table 16 gray average

Table 17 standard deviation

Table 18 Clustering Tendency

Table 19 homogeneity degree

A-1_1_H	A-1_2_H	A-1_3_H	A-1_4_H	A-1_5_H	A-1_6_H
						0.729353	0.85896	0.742591	0.811983	0.707436	0.820406
A-2_1_H	A-2_2_H	A-2_3_H	A-2_4_H	A-2_5_H	A-2_6_H
						0.673075	0.832321	0.804619	0.818736	0.765196	0.819104
A-3_1_H	A-3_2_H	A-3_3_H	A-3_4_H	A-3_5_H	A-3_6_H
						0.702889	0.770329	0.731664	0.812325	0.750113	0.823354
A-4_1_H	A-4_2_H	A-4_3_H	A-4_4_H	A-4_5_H	A-4_6_H
						0.630094	0.809826	0.765024	0.750887	0.722146	0.833379
A-5_1_H	A-5_2_H	A-5_3_H	A-5_4_H	A-5_5_H	A-5_6_H
						0.784226	0.802753	0.828898	0.824906	0.855482	0.832783
A-6_1_H	A-6_2_H	A-6_3_H	A-6_4_H	A-6_5_H	A-6_6_H
						0.860491	0.882292	0.902136	0.888263	0.899928	0.892365
A-7_1_H	A-7_2_H	A-7_3_H	A-7_4_H	A-7_5_H	A-7_6_H
						0.860491	0.872411	0.897529	0.897329	0.909925	0.913109
A-8_1_H	A-8_2_H	A-8_3_H	A-8_4_H	A-8_5_H	A-8_6_H
						0.720064	0.846891	0.845386	0.786799	0.819207	0.776603
B-1_1_H	B-1_2_H	B-1_3_H	B-1_4_H	B-1_5_H	B-1_6_H
						0.421003	0.63488	0.678265	0.758653	0.853534	0.902955
B-2_1_H	B-2_2_H	B-2_3_H	B-2_4_H	B-2_5_H	B-2_6_H
						0.413449	0.723771	0.737913	0.847828	0.900318	0.919516
B-3_1_H	B-3_2_H	B-3_3_H	B-3_4_H	B-3_5_H	B-3_6_H
						0.404989	0.67885	0.73714	0.734893	0.901268	0.88582
B-4_1_H	B-4_2_H	B-4_3_H	B-4_4_H	B-4_5_H	B-4_6_H
						0.379105	0.759241	0.706057	0.764712	0.862515	0.842974
B-5_1_H	B-5_2_H	B-5_3_H	B-5_4_H	B-5_5_H	B-5_6_H
						0.386228	0.803288	0.741632	0.777376	0.934154	0.900476
B-6_1_H	B-6_2_H	B-6_3_H	B-6_4_H	B-6_5_H	B-6_6_H
						0.492837	0.64838	0.84093	0.804451	0.934574	0.924924

B-7_1_H	B-7_2_H	B-7_3_H	B-7_4_H	B-7_5_H	B-7_6_H
						0.546176	0.825809	0.848521	0.875017	0.923975	0.925081
B-8_1_H	B-8_2_H	B-8_3_H	B-8_4_H	B-8_5_H	B-8_6_H
						0.341506	0.648599	0.822798	0.637463	0.89822	0.891833

Table 20 maximum probability

Table 21 energy

A-1_1_ energy	A-1_2_ energy	A-1_3_ energy	A-1_4_ energy	A-1_5_ energy	A-1_6_ energy
						0.329817	0.429658	0.269169	0.357082	0.234291	0.360617
A-2_1_ energy	A-2_2_ energy	A-2_3_ energy	A-2_4_ energy	A-2_5_ energy	A-2_6_ energy
						0.252431	0.346633	0.319042	0.347412	0.283623	0.357081
A-3_1_ energy	A-3_2_ energy	A-3_3_ energy	A-3_4_ energy	A-3_5_ energy	A-3_6_ energy
						0.253393	0.27921	0.260922	0.328612	0.279507	0.329821
A-4_1_ energy	A-4_2_ energy	A-4_3_ energy	A-4_4_ energy	A-4_5_ energy	A-4_6_ energy
						0.262432	0.326571	0.281119	0.268873	0.252395	0.329804
A-5_1_ energy	A-5_2_ energy	A-5_3_ energy	A-5_4_ energy	A-5_5_ energy	A-5_6_ energy
						0.301237	0.305192	0.347685	0.334801	0.378429	0.328888
A-6_1_ energy	A-6_2_ energy	A-6_3_ energy	A-6_4_ energy	A-6_5_ energy	A-6_6_ energy
						0.380532	0.397709	0.422509	0.390428	0.431073	0.399358
A-7_1_ energy	A-7_2_ energy	A-7_3_ energy	A-7_4_ energy	A-7_5_ energy	A-7_6_ energy
						0.380532	0.385016	0.413217	0.421572	0.4422	0.437127
A-8_1_ energy	A-8_2_ energy	A-8_3_ energy	A-8_4_ energy	A-8_5_ energy	A-8_6_ energy
						0.235615	0.358153	0.365301	0.285675	0.385935	0.285943
B-1_1_ energy	B-1_2_ energy	B-1_3_ energy	B-1_4_ energy	B-1_5_ energy	B-1_6_ energy
						0.040559	0.152371	0.145938	0.322172	0.562941	0.703012
B-2_1_ energy	B-2_2_ energy	B-2_3_ energy	B-2_4_ energy	B-2_5_ energy	B-2_6_ energy
						0.047743	0.224751	0.327096	0.587206	0.694283	0.746295
B-3_1_ energy	B-3_2_ energy	B-3_3_ energy	B-3_4_ energy	B-3_5_ energy	B-3_6_ energy
						0.036529	0.199446	0.271257	0.326123	0.673829	0.615674
B-4_1_ energy	B-4_2_ energy	B-4_3_ energy	B-4_4_ energy	B-4_5_ energy	B-4_6_ energy
						0.02904	0.37342	0.237587	0.371034	0.590966	0.509126

B-5_1_ energy	B-5_2_ energy	B-5_3_ energy	B-5_4_ energy	B-5_5_ energy	B-5_6_ energy
						0.028061	0.46962	0.284735	0.343905	0.79342	0.696714
B-6_1_ energy	B-6_2_ energy	B-6_3_ energy	B-6_4_ energy	B-6_5_ energy	B-6_6_ energy
						0.065538	0.159744	0.563571	0.382843	0.794607	0.754498
B-7_1_ energy	B-7_2_ energy	B-7_3_ energy	B-7_4_ energy	B-7_5_ energy	B-7_6_ energy
						0.113649	0.562378	0.457991	0.611939	0.732534	0.757845
B-8_1_ energy	B-8_2_ energy	B-8_3_ energy	B-8_4_ energy	B-8_5_ energy	B-8_6_ energy
						0.027273	0.133006	0.501218	0.126851	0.585713	0.642467

Table 22 moment of inertia

Table 23 unfavourable balance distance

A-1_1_I_D_M	A-1_2_I_D_M	A-1_3_I_D_M	A-1_4_I_D_M	A-1_5_I_D_M	A-1_6_I_D_M
						0.712587	0.849413	0.720855	0.794639	0.679748	0.804419
A-2_1_I_D_M	A-2_2_I_D_M	A-2_3_I_D_M	A-2_4_I_D_M	A-2_5_I_D_M	A-2_6_I_D_M
						0.642687	0.818283	0.785718	0.803443	0.743536	0.803096
A-3_1_I_D_M	A-3_2_I_D_M	A-3_3_I_D_M	A-3_4_I_D_M	A-3_5_I_D_M	A-3_6_I_D_M
						0.687918	0.752437	0.708439	0.797982	0.727894	0.809125
A-4_1_I_D_M	A-4_2_I_D_M	A-4_3_I_D_M	A-4_4_I_D_M	A-4_5_I_D_M	A-4_6_I_D_M
						0.603872	0.796356	0.74549	0.729865	0.697074	0.819374
A-5_1_I_D_M	A-5_2_I_D_M	A-5_3_I_D_M	A-5_4_I_D_M	A-5_5_I_D_M	A-5_6_I_D_M
						0.772473	0.785752	0.813567	0.811672	0.844024	0.819457
A-6_1_I_D_M	A-6_2_I_D_M	A-6_3_I_D_M	A-6_4_I_D_M	A-6_5_I_D_M	A-6_6_I_D_M
						0.851849	0.87395	0.894326	0.877756	0.891504	0.883767
A-7_1_I_D_M	A-7_2_I_D_M	A-7_3_I_D_M	A-7_4_I_D_M	A-7_5_I_D_M	A-7_6_I_D_M
						0.851849	0.860903	0.888111	0.88889	0.902253	0.906185
A-8_1_I_D_M	A-8_2_I_D_M	A-8_3_I_D_M	A-8_4_I_D_M	A-8_5_I_D_M	A-8_6_I_D_M
						0.698492	0.837275	0.832835	0.767781	0.804973	0.756187
B-1_1_I_D_M	B-1_2_I_D_M	B-1_3_I_D_M	B-1_4_I_D_M	B-1_5_I_D_M	B-1_6_I_D_M
						0.345231	0.598857	0.655871	0.737716	0.838656	0.893359
B-2_1_I_D_M	B-2_2_I_D_M	B-2_3_I_D_M	B-2_4_I_D_M	B-2_5_I_D_M	B-2_6_I_D_M
						0.345887	0.702856	0.70873	0.829426	0.889851	0.911388

B-3_1_I_D_M	B-3_2_I_D_M	B-3_3_I_D_M	B-3_4_I_D_M	B-3_5_I_D_M	B-3_6_I_D_M
						0.3318	0.6468	0.713197	0.705259	0.893005	0.877113
B-4_1_I_D_M	B-4_2_I_D_M	B-4_3_I_D_M	B-4_4_I_D_M	B-4_5_I_D_M	B-4_6_I_D_M
						0.292172	0.732955	0.678273	0.739201	0.848489	0.829223
B-5_1_I_D_M	B-5_2_I_D_M	B-5_3_I_D_M	B-5_4_I_D_M	B-5_5_I_D_M	B-5_6_I_D_M
						0.308495	0.779828	0.724758	0.76087	0.926433	0.890352
B-6_1_I_D_M	B-6_2_I_D_M	B-6_3_I_D_M	B-6_4_I_D_M	B-6_5_I_D_M	B-6_6_I_D_M
						0.443182	0.620292	0.822765	0.788644	0.927426	0.917631
B-7_1_I_D_M	B-7_2_I_D_M	B-7_3_I_D_M	B-7_4_I_D_M	B-7_5_I_D_M	B-7_6_I_D_M
						0.496028	0.80542	0.837661	0.86274	0.917396	0.917402
B-8_1_I_D_M	B-8_2_I_D_M	B-8_3_I_D_M	B-8_4_I_D_M	B-8_5_I_D_M	B-8_6_I_D_M
						0.247864	0.621626	0.803763	0.608389	0.892866	0.881686

Table 24 entropy entropy

A-1_1_ entropy	A-1_2_ entropy	A-1_3_ entropy	A-1_4_ entropy	A-1_5_ entropy	A-1_6_ entropy
						0.147365	0.249842	0.809925	0.674705	0.991099	0.707431
A-2_1_ entropy	A-2_2_ entropy	A-2_3_ entropy	A-2_4_ entropy	A-2_5_ entropy	A-2_6_ entropy
						0.257654	0.467325	0.744689	0.634534	0.841402	0.719484
A-3_1_ entropy	A-3_2_ entropy	A-3_3_ entropy	A-3_4_ entropy	A-3_5_ entropy	A-3_6_ entropy
						0.257654	0.78344	0.941854	0.577694	0.86642	0.638637
A-4_1_ entropy	A-4_2_ entropy	A-4_3_ entropy	A-4_4_ entropy	A-4_5_ entropy	A-4_6_ entropy
						0.5331	0.435041	0.892007	0.737686	0.926049	0.527086
A-5_1_ entropy	A-5_2_ entropy	A-5_3_ entropy	A-5_4_ entropy	A-5_5_ entropy	A-5_6_ entropy
						0.213101	0.673818	0.609655	0.569882	0.483188	0.542628
A-6_1_ entropy	A-6_2_ entropy	A-6_3_ entropy	A-6_4_ entropy	A-6_5_ entropy	A-6_6_ entropy
						0.147365	0.402757	0.533242	0.542628	0.402109	0.370703
A-7_1_ entropy	A-7_2_ entropy	A-7_3_ entropy	A-7_4_ entropy	A-7_5_ entropy	A-7_6_ entropy

0.147365	0.449194	0.504984	0.419245	0.339453	0.314577
						A-8_1_ entropy	A-8_2_ entropy	A-8_3_ entropy	A-8_4_ entropy	A-8_5_ entropy	A-8_6_ entropy
0.459046	0.485728	0.599459	0.827273	0.63416	0.862612
						B-1_1_ entropy	B-1_2_ entropy	B-1_3_ entropy	B-1_4_ entropy	B-1_5_ entropy	B-1_6_ entropy
1.135293	1.372649	0.989272	0.907689	0.623399	0.407773
						B-2_1_ entropy	B-2_2_ entropy	B-2_3_ entropy	B-2_4_ entropy	B-2_5_ entropy	B-2_6_ entropy
0.960201	1.047255	1.102955	0.80623	0.460222	0.339453
						B-3_1_ entropy	B-3_2_ entropy	B-3_3_ entropy	B-3_4_ entropy	B-3_5_ entropy	B-3_6_ entropy
1.226785	1.259216	0.944822	1.048946	0.364806	0.394297
						B-4_1_ entropy	B-4_2_ entropy	B-4_3_ entropy	B-4_4_ entropy	B-4_5_ entropy	B-4_6_ entropy
1.276276	1.092116	1.066606	1.01884	0.502752	0.594869
						B-5_1_ entropy	B-5_2_ entropy	B-5_3_ entropy	B-5_4_ entropy	B-5_5_ entropy	B-5_6_ entropy
1.311138	1.098337	0.707474	0.769707	0.306765	0.404825
						B-6_1_ entropy	B-6_2_ entropy	B-6_3_ entropy	B-6_4_ entropy	B-6_5_ entropy	B-6_6_ entropy
0.967183	1.21674	0.765465	0.833332	0.302827	0.361857
						B-7_1_ entropy	B-7_2_ entropy	B-7_3_ entropy	B-7_4_ entropy	B-7_5_ entropy	B-7_6_ entropy
1.010607	1.027496	0.585507	0.662255	0.277905	0.354045
						B-8_1_ entropy	B-8_2_ entropy	B-8_3_ entropy	B-8_4_ entropy	B-8_5_ entropy	B-8_6_ entropy
1.25816	1.235578	0.807947	0.975353	0.25531	0.599459

Table 25 correlativity

A-1_1_C	A-1_2_C	A-1_3_C	A-1_4_C	A-1_5_C	A-1_6_C
						0.453488	0.691889	0.611768	0.7148	0.62426	0.732309
A-2_1_C	A-2_2_C	A-2_3_C	A-2_4_C	A-2_5_C	A-2_6_C
						0.459653	0.735835	0.733302	0.715312	0.697015	0.732849
A-3_1_C	A-3_2_C	A-3_3_C	A-3_4_C	A-3_5_C	A-3_6_C
						0.495842	0.652304	0.653717	0.67898	0.656804	0.746605

A-4_1_C	A-4_2_C	A-4_3_C	A-4_4_C	A-4_5_C	A-4_6_C
						0.35345	0.651618	0.666332	0.599482	0.629691	0.742986
A-5_1_C	A-5_2_C	A-5_3_C	A-5_4_C	A-5_5_C	A-5_6_C
						0.595731	0.712844	0.721926	0.704375	0.756268	0.731947
A-6_1_C	A-6_2_C	A-6_3_C	A-6_4_C	A-6_5_C	A-6_6_C
						0.713227	0.774956	0.844911	0.837104	0.84023	0.840448
A-7_1_C	A-7_2_C	A-7_3_C	A-7_4_C	A-7_5_C	A-7_6_C
						0.713227	0.80898	0.854349	0.820523	0.864554	0.865475
A-8_1_C	A-8_2_C	A-8_3_C	A-8_4_C	A-8_5_C	A-8_6_C
						0.583363	0.732909	0.745791	0.723227	0.706181	0.70656
B-1_1_C	B-1_2_C	B-1_3_C	B-1_4_C	B-1_5_C	B-1_6_C
						-0.06096	0.02836	0.181045	0.07172	-0.04784	-0.17716
B-2_1_C	B-2_2_C	B-2_3_C	B-2_4_C	B-2_5_C	B-2_6_C
						-0.11077	-0.02076	-0.04274	-0.16053	-0.17078	-0.18233
B-3_1_C	B-3_2_C	B-3_3_C	B-3_4_C	B-3_5_C	B-3_6_C
						-0.08765	-0.03503	-0.05094	-0.15471	-0.0704	-0.15323
B-4_1_C	B-4_2_C	B-4_3_C	B-4_4_C	B-4_5_C	B-4_6_C
						-0.0882	-0.05643	0.025384	0.173975	-0.14326	-0.09191
B-5_1_C	B-5_2_C	B-5_3_C	B-5_4_C	B-5_5_C	B-5_6_C
						-0.12153	-0.12012	-0.12315	0.022757	-0.19152	-0.24968
B-6_1_C	B-6_2_C	B-6_3_C	B-6_4_C	B-6_5_C	B-6_6_C
						0.136108	-0.03391	-0.20318	0.027926	-0.11888	0.034783
B-7_1_C	B-7_2_C	B-7_3_C	B-7_4_C	B-7_5_C	B-7_6_C
						0.064315	-0.18657	-0.19499	-0.2193	-0.12311	-0.09605
B-8_1_C	B-8_2_C	B-8_3_C	B-8_4_C	B-8_5_C	B-8_6_C
						0.185817	-0.05458	-0.06841	0.120048	0.005032	-0.03155

Table 26 and mean

The mean that table 27 is poor

A-1_1_D-m	A-1_2_D-m	A-1_3_D-m	A-1_4_D-m	A-1_5_D-m	A-1_6_D-m
						4.330357	2.138244	2.994348	2.142065	2.835198	1.997181
A-2_1_D-m	A-2_2_D-m	A-2_3_D-m	A-2_4_D-m	A-2_5_D-m	A-2_6_D-m

4.200893	2.104315	2.19344	2.239163	2.377385	2.019957
						A-3_1_D-m	A-3_2_D-m	A-3_3_D-m	A-3_4_D-m	A-3_5_D-m	A-3_6_D-m
3.879464	2.75506	2.739019	2.507524	2.533449	2.016009
						A-4_1_D-m	A-4_2_D-m	A-4_3_D-m	A-4_4_D-m	A-4_5_D-m	A-4_6_D-m
4.889881	2.667708	2.62626	3.174323	2.70422	2.090096
						A-5_1_D-m	A-5_2_D-m	A-5_3_D-m	A-5_4_D-m	A-5_5_D-m	A-5_6_D-m
3.087798	2.303869	2.19344	2.302779	1.84492	2.148413
						A-6_1_D-m	A-6_2_D-m	A-6_3_D-m	A-6_4_D-m	A-6_5_D-m	A-6_6_D-m
2.232143	1.689137	1.21389	1.342886	1.207465	1.282544
						A-7_1_D-m	A-7_2_D-m	A-7_3_D-m	A-7_4_D-m	A-7_5_D-m	A-7_6_D-m
2.232143	1.553125	1.202153	1.374208	1.05587	1.044899
						A-8_1_D-m	A-8_2_D-m	A-8_3_D-m	A-8_4_D-m	A-8_5_D-m	A-8_6_D-m
3.361607	2.006845	1.915359	2.28845	2.008073	2.340365
						B-1_1_D-m	B-1_2_D-m	B-1_3_D-m	B-1_4_D-m	B-1_5_D-m	B-1_6_D-m
3.110119	1.439583	1.003132	0.857143	0.535334	0.376744
						B-2_1_D-m	B-2_2_D-m	B-2_3_D-m	B-2_4_D-m	B-2_5_D-m	B-2_6_D-m
3.212798	0.959375	1.100518	0.656178	0.374198	0.299769
						B-3_1_D-m	B-3_2_D-m	B-3_3_D-m	B-3_4_D-m	B-3_5_D-m	B-3_6_D-m
3.43006	1.245685	0.985383	1.114523	0.338087	0.364959
						B-4_1_D-m	B-4_2_D-m	B-4_3_D-m	B-4_4_D-m	B-4_5_D-m	B-4_6_D-m
3.641369	0.976935	1.210397	0.853183	0.549405	0.552794
						B-5_1_D-m	B-5_2_D-m	B-5_3_D-m	B-5_4_D-m	B-5_5_D-m	B-5_6_D-m
3.455357	0.912946	0.920111	0.751692	0.269593	0.430039
						B-6_1_D-m	B-6_2_D-m	B-6_3_D-m	B-6_4_D-m	B-6_5_D-m	B-6_6_D-m
2.845238	1.385119	0.744168	0.758785	0.265551	0.302827
						B-7_1_D-m	B-7_2_D-m	B-7_3_D-m	B-7_4_D-m	B-7_5_D-m	B-7_6_D-m
2.665179	1.012649	0.705357	0.521025	0.317952	0.29919
						B-8_1_D-m	B-8_2_D-m	B-8_3_D-m	B-8_4_D-m	B-8_5_D-m	B-8_6_D-m

3.28125

1.248958

0.714754

1.215978

0.300066

0.368643

Table 28 and entropy

A-1_1_S-E	A-1_2_S-E	A-1_3_S-E	A-1_4_S-E	A-1_5_S-E	A-1_6_S-E
						0.612069	1.011892	2.75211	2.559395	2.653717	2.357032
A-2_1_S-E	A-2_2_S-E	A-2_3_S-E	A-2_4_S-E	A-2_5_S-E	A-2_6_S-E
						0.947376	1.724922	2.397386	2.264199	2.474822	2.136188
A-3_1_S-E	A-3_2_S-E	A-3_3_S-E	A-3_4_S-E	A-3_5_S-E	A-3_6_S-E
						1.011892	2.539099	2.962306	1.771709	2.437167	2.382927
A-4_1_S-E	A-4_2_S-E	A-4_3_S-E	A-4_4_S-E	A-4_5_S-E	A-4_6_S-E
						2.028047	1.531374	2.856791	2.269515	2.485508	1.92766
A-5_1_S-E	A-5_2_S-E	A-5_3_S-E	A-5_4_S-E	A-5_5_S-E	A-5_6_S-E
						0.748327	2.132039	2.136188	2.055342	1.836225	2.264199
A-6_1_S-E	A-6_2_S-E	A-6_3_S-E	A-6_4_S-E	A-6_5_S-E	A-6_6_S-E
						0.612069	1.571539	1.838792	2.058986	1.853822	1.836225
A-7_1_S-E	A-7_2_S-E	A-7_3_S-E	A-7_4_S-E	A-7_5_S-E	A-7_6_S-E
						0.612069	1.911607	1.905091	1.727867	1.942689	1.612649
A-8_1_S-E	A-8_2_S-E	A-8_3_S-E	A-8_4_S-E	A-8_5_S-E	A-8_6_S-E
						1.914239	1.671989	2.242653	2.382927	2.41185	2.593117
B-1_1_S-E	B-1_2_S-E	B-1_3_S-E	B-1_4_S-E	B-1_5_S-E	B-1_6_S-E
						2.943925	3.07012	2.694398	2.619668	2.056985	1.938491
B-2_1_S-E	B-2_2_S-E	B-2_3_S-E	B-2_4_S-E	B-2_5_S-E	B-2_6_S-E
						2.861919	2.483775	2.804868	2.191245	1.778759	1.726835
B-3_1_S-E	B-3_2_S-E	B-3_3_S-E	B-3_4_S-E	B-3_5_S-E	B-3_6_S-E
						3.186176	2.732559	2.705314	2.602177	1.726835	1.726835
B-4_1_S-E	B-4_2_S-E	B-4_3_S-E	B-4_4_S-E	B-4_5_S-E	B-4_6_S-E
						3.366478	2.684184	2.921308	2.722804	2.083337	2.235981
B-5_1_S-E	B-5_2_S-E	B-5_3_S-E	B-5_4_S-E	B-5_5_S-E	B-5_6_S-E

3.097308	2.721316	2.339401	2.260333	1.656175	1.911607
						B-6_1_S-E	B-6_2_S-E	B-6_3_S-E	B-6_4_S-E	B-6_5_S-E	B-6_6_S-E
3.081967	2.726498	2.359926	2.417582	1.637968	1.809459
						B-7_1_S-E	B-7_2_S-E	B-7_3_S-E	B-7_4_S-E	B-7_5_S-E	B-7_6_S-E
3.18791	2.696965	1.954305	2.050971	1.612649	1.72128
						B-8_1_S-E	B-8_2_S-E	B-8_3_S-E	B-8_4_S-E	B-8_5_S-E	B-8_6_S-E
3.471159	2.899189	2.437167	2.71465	1.466858	1.825372

The entropy that table 29 is poor

A-1_1_D-E	A-1_2_D-E	A-1_3_D-E	A-1_4_D-E	A-1_5_D-E	A-1_6_D-E
						0.668564	1.311278	3	2.899397	2.875	2.477217
A-2_1_D-E	A-2_2_D-E	A-2_3_D-E	A-2_4_D-E	A-2_5_D-E	A-2_6_D-E
						1.311278	1.621641	2.82782	2.655639	3.030639	2.423795
A-3_1_D-E	A-3_2_D-E	A-3_3_D-E	A-3_4_D-E	A-3_5_D-E	A-3_6_D-E
						1.311278	3.125	2.95282	2.483459	2.521782	2.305037
A-4_1_D-E	A-4_2_D-E	A-4_3_D-E	A-4_4_D-E	A-4_5_D-E	A-4_6_D-E
						2.271782	2	3.125	2.858459	2.5	2.125
A-5_1_D-E	A-5_2_D-E	A-5_3_D-E	A-5_4_D-E	A-5_5_D-E	A-5_6_D-E
						0.993393	2.655639	2.483459	2.57782	2.555037	2.436278
A-6_1_D-E	A-6_2_D-E	A-6_3_D-E	A-6_4_D-E	A-6_5_D-E	A-6_6_D-E
						0.668564	2.07782	2.298795	2.649397	2.405639	2.280639
A-7_1_D-E	A-7_2_D-E	A-7_3_D-E	A-7_4_D-E	A-7_5_D-E	A-7_6_D-E
						0.668564	2.349602	2.375	2.030639	1.794737	2.07782
A-8_1_D-E	A-8_2_D-E	A-8_3_D-E	A-8_4_D-E	A-8_5_D-E	A-8_6_D-E
						2.375	2.07782	2.75	2.75	2.780639	3.030639
B-1_1_D-E	B-1_2_D-E	B-1_3_D-E	B-1_4_D-E	B-1_5_D-E	B-1_6_D-E
						3.155639	3.405639	2.521782	2.771782	2.521782	2.349602

B-2_1_D-E	B-2_2_D-E	B-2_3_D-E	B-2_4_D-E	B-2_5_D-E	B-2_6_D-E
						3.32782	2.774397	3.25	2.474602	2.530639	2.20282
B-3_1_D-E	B-3_2_D-E	B-3_3_D-E	B-3_4_D-E	B-3_5_D-E	B-3_6_D-E
						3.45282	3.108459	2.875	2.95282	2.07782	2.20282
B-4_1_D-E	B-4_2_D-E	B-4_3_D-E	B-4_4_D-E	B-4_5_D-E	B-4_6_D-E
						3.45282	3	3.280639	2.646782	2.655639	2.608459
B-5_1_D-E	B-5_2_D-E	B-5_3_D-E	B-5_4_D-E	B-5_5_D-E	B-5_6_D-E
						3.20282	3.024397	2.82782	2.477217	2.20282	2.548795
B-6_1_D-E	B-6_2_D-E	B-6_3_D-E	B-6_4_D-E	B-6_5_D-E	B-6_6_D-E
						3.108459	2.95282	2.852217	2.852217	2.07782	2.271782
B-7_1_D-E	B-7_2_D-E	B-7_3_D-E	B-7_4_D-E	B-7_5_D-E	B-7_6_D-E
						3.625	2.95282	2.20282	2.436278	2.305037	2.25
B-8_1_D-E	B-8_2_D-E	B-8_3_D-E	B-8_4_D-E	B-8_5_D-E	B-8_6_D-E
						3.024397	3.07782	2.75	3	1.919737	2.375

Claims (3)

1. the various dimensions texture extracting method based on crowd's brain nuclear magnetic resonance image, is characterized in that: deployment area growth method by the Region Segmentation needing in crowd's brain nuclear magnetic resonance image out.

2. extracting method according to claim 1, it is characterized in that: adopt Curvelet transform method to extract the textural characteristics parameter in the region of needs, wherein said crowd comprises Alzheimer disease people colony, mild cognitive impairment patient population and normal aging people colony, the textural characteristics parameter in the region of described needs comprises entropy, gray average, correlativity, energy, homogeneity degree, variance, maximum probability, unfavourable balance distance, Clustering Tendency, contrast, with average, poor average, with entropy, poor entropy, the region of described needs comprises entorhinal cortex and two regions of hippocampus.

3. extracting method according to claim 1, it is characterized in that: adopt Contourlet transform method to extract the textural characteristics parameter in the region of needs, wherein said crowd comprises Alzheimer disease people colony, mild cognitive impairment patient population and normal aging people colony, the textural characteristics parameter in the region of described needs comprises entropy, gray average, correlativity, energy, homogeneity degree, variance, maximum probability, unfavourable balance distance, Clustering Tendency, contrast, with average, poor average, with entropy, poor entropy, the region of described needs comprises entorhinal cortex and two regions of hippocampus.

Images