CN101699515A - Multi-elite immune quantum clustering-based medical image segmenting system and multi-elite immune quantum clustering-based medical image segmenting method - Google Patents

Abstract

The invention discloses a multi-elite immune quantum clustering-based medical image segmenting system and a multi-elite immune quantum clustering-based medical image segmenting method, which relate to the technical field of image processing. The system comprises a preprocessing module, a data preparing module, a data clustering module and a segmentation result output module. The process for segmenting a medical image by the modules comprises the following steps: 1) preprocessing the medical image to be segmented; 2) coding antibodies and initializing an antibody population; 3) calculating antibody affinity, and dividing the antibody population into an elite population and a general population; 4) designing different multi-elite immune optimization operators for the elite population and the general population respectively, and performing a cloning operation, a cloud mutation operation, an all-interference recombination operation, a selecting operation and a hypercube interlace operation orderly; and 5) outputting a medical image segmentation result. The multi-elite immune quantum clustering-based medical image segmenting system and the multi-elite immune quantum clustering-based medical image segmenting method can effectively segment the medical image which contains large-scale data volume, has an accurate and precise segmentation result, and can be used for auxiliary diagnosis of the medical image and pathogenesis research.

Description

Medical image segmenting system and dividing method based on many elite's immunity quantum clusterings

Technical field

The invention belongs to technical field of image processing, relate to medical image and cut apart, can be used for medical image auxiliary diagnosis and research pathogenesis.

Background technology

Along with rapid development of computer technology, more and more important effect is being brought into play in computer-aided diagnosis in clinical medicine.Traditional judges pathogenetic method according to naked eyes, is replaced by technology such as X ray, CT and MRI gradually, is greatly promoting the development that Medical Imaging and disease research are learned.

Image segmentation is divided into several significant parts according to certain homogeneity or conforming principle with image exactly, interested object is extracted from the ten minutes complicated background, so that further analyze.Medical image is owing to be subjected to little, the image blurring inhomogeneous and noise of its gray difference and factor affecting such as the pathology classification is many, cause the auxiliary diagnosis of medical image become one very complicated and press for the problem of solution.To cutting apart of medical image, help the institutional framework that human body is meticulous to identify, and, further diseased region is positioned in conjunction with human anatomy priori and pathological correlative study achievement, be used for medical diagnosis on disease and treatment.

The method of using cluster is cut apart medical image, be that the discontinuous part of one or more features in the medical image is divided into a sub regions separately, raw information is converted into compact more form, makes higher level graphical analysis and understanding become possibility.Cluster is meant under without any the priori situation about sample, utilizes the method research of mathematics and handles the classification of special object, do not have one the sample of classification mark to be divided into several subclass according to certain criterion.

As a kind of quantum clustering based on the nonparametric clustering technique of dividing, it can overcome traditional clustering method initial value and noise-sensitive, cluster classification number are wanted defective such as in advance given.But the method that quantum clustering descends by gradient is carried out iteration and is easy to be absorbed in local extremum, and simultaneously, iteration speed has limited it at the large-scale dataset especially application in image segmentation field slowly.Although also there are some improvement technology, for example, the improvement estimated of adjusting the distance, improvement that scale parameter is estimated etc. all fail fundamentally to solve above bottleneck problem.

Summary of the invention

The objective of the invention is to overcome the deficiency of above-mentioned prior art, a kind of medical image segmenting system and dividing method based on many elite's immunity quantum clusterings proposed, many elite immune optimization theory is combined with quantum clustering, can cut apart the medical image that comprises the large-scale data amount effectively, the organization of human body tissue is carried out accurate and meticulous location, improve the medical imaging diagnosis level.

For achieving the above object, the invention provides medical image segmenting system, comprising based on many elite's immunity quantum clusterings:

The medical image pretreatment module, convert the medical image of rgb format to gray level image, and carry out the histogram equalization enhancement process, and the gradation of image value after handling according to pixels to be put from top to bottom, order from left to right forms a line and is input to the medical image data preparation module;

The medical image data preparation module is carried out cluster centre coding to the gray value data sample of input, forms the antibody of many elite's immunity quantum clusterings, and picked at random k antibody is as the initial antibodies population, k be medical image cut apart the classification number;

Medical image data cluster module is used many elite's immunity quantum clustering antagonist populations and is carried out cluster, and exports cluster result to medical image segmentation result output module;

Medical image segmentation result output module turns back to the cluster result label of importing in the gray level image, exports final medical image segmentation result.

Described medical image data cluster module comprises:

Antibody affinity degree function design submodule, the schrodinger equation that is used for finding the solution quantum mechanics obtains the potential-energy function computing formula, and this potential-energy function minimal point is corresponding with cluster centre, according to cluster centre calculating affinity degree functional value;

The sub-population of many elite is divided submodule, according to affinity degree ranking results, antibody population is divided into sub-population of elite and common sub-population;

Many elite immune optimization operation operator design submodule is used for carrying out successively the antibody cloning operation, the sub-population of elite is adopted the cloud mutation operation, disturbs the reorganization operation and utilize the hypercube interlace operation that sub-population of elite and common sub-population are carried out coevolution common sub-population employing is absolutely dry.

For achieving the above object, the invention provides medical image dividing method, comprise the steps: based on many elite's immunity quantum clusterings

(1) convert medical image to be split to gray level image, utilize histogram equalization method degree of comparing to strengthen, the gradation of image value after handling is according to pixels put from top to bottom, order from left to right forms a line;

(2) above-mentioned gray value data is carried out the cluster centre coding, form the antibody of many elite's immunity quantum clusterings, each antibody length is N * k, wherein, first length is that the field of N is represented first cluster centre, and the field that second length is N is represented second cluster centre, and the rest may be inferred;

(3) from the antibody behind the coding picked at random k antibody as the initial antibodies population, k be medical image cut apart the classification number;

(4) the affinity degree value of calculating antibody population:

4a) two antibody sample point x of definition _iAnd x _jBetween distance function D _Ij=|| x _i-x _j||, find the solution schrodinger equation

The computing formula that obtains potential-energy function V is:

$V=E-\frac{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">d</figure-callout>}}{2}+\frac{1}{{2\mathrm{\&sigma;}}^2}\frac{{\mathrm{\&Sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="j\; D\; ij"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">j</figure-callout>}}{D}_{\mathrm{ij}}^{}{}_2^{}\exp (-\frac{{\mathrm{<figure-callout\; id="2"\; label="D\; ij"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">D</figure-callout>}}_{\mathrm{ij}}^{}}{{2\mathrm{\&sigma;}}^2})}{{\mathrm{\&Sigma;}}_j\exp (-\frac{{\mathrm{<figure-callout\; id="2"\; label="D\; ij"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">D</figure-callout>}}_{\mathrm{ij}}^{}}{{2\mathrm{\&sigma;}}^2})}$

Wherein, H is a Hamiltonian,

E is the energy feature value of Hamiltonian, and d is the intrinsic dimensionality of input data, and x is the input sample of data point, and σ is a scale parameter,

Be Laplace operator;

4b) determine K cluster centre { c according to the minimal point of the potential-energy function value that calculates _i, i=1 ... K}, and according to sample point and all kinds of division set Q of the nearest principle acquisition of each cluster centre Euclidean distance _i, then affinity degree function calculation formula is:

$f=1/(1+\underset{i=1}{\overset{K}{\mathrm{\&Sigma;}}}\underset{x_j\&Element;O_i}{\mathrm{\&Sigma;}}||x_j-c_i|\left|\right)$

Wherein, || || for asking for the operational symbol of Euclidean distance;

(5) calculate the affinity degree value of all antibody in the antibody population after, they are sorted from high to low, pM antibody getting the front is as the sub-population of elite, remainder is as common sub-population, wherein, the ratio value of p for setting, M is the population size;

(6) to the sub-population of elite and common sub-population according to the clone's scale N that sets _cCarry out clone operations;

(7) after the clone operations, the sub-population of elite is carried out the cloud mutation operation, common sub-population is carried out the absolutely dry reorganization operation of disturbing, concrete steps are as follows:

7a) produce several r between one (0,1) at random, if satisfy r＜P _m, P _mBe the variation probability of setting, then the sub-population antibody of elite after the variation is Wherein, Ex is the preceding antibody of variation, and En=En+Herandn, En=0.1 σ, σ are the standard deviation of each dimension data variable, and He=0.1En, randn represent to satisfy the random number of standardized normal distribution, μ=P _Max-(P _Max-P _Min) (f _Max-f)/(f _Max+ f _Min) expression degree of certainty value, wherein, f, f _MaxAnd f _MinBe respectively affinity degree value, affinity degree maximal value and the affinity degree minimum value of each iteration, P _MaxAnd P _MinBe respectively the degree of certainty maximal value and the degree of certainty minimum value of setting;

7b) for common sub-population, all antibody all participate in reorganization, carry out permutation and combination according to the diagonal line rule;

(8) the sub-population of antibody after operating through clone operations, mutation operation and reorganization, select outstanding antibody to form sub-population of new elite and common sub-population according to probability, the affinity degree mxm. of the sub-population of elite that this is new is fitb1, and the affinity degree mxm. of new common sub-population is fitb2;

(9) the affinity degree mxm. with the affinity degree mxm. of the new sub-population of elite and new common sub-population compares, if fitb1 〉=fitb2, pM just that the antibody of the sub-population of elite and common sub-population affinity degree is minimum antibody carries out hypercube and intersects; If fitb1＜fitb2, pM just that common sub-population affinity degree is the highest antibody carries out hypercube with the sub-population antibody of elite and intersects;

(10) through after the interlace operation, return step (4) and carry out iteration optimization again, repeat N _MaxInferior;

(11) will be through N _MaxThe affinity degree mxm. fitb2 of affinity degree mxm. fitb1 of the sub-population of elite that inferior iteration obtains at last and common sub-population compares again, with the final cluster centre of the pairing antibody representative of the higher value after the comparison, and be divided in the different classifications according to this cluster centre each pixel with medical image, obtain final segmentation result;

The present invention has the following advantages compared with prior art:

1) the present invention has introduced the many elite immunologic process with biology immune mechanism, can effectively overcome the defective that is absorbed in local extremum in the quantum clustering iterative process easily;

2) the present invention can effectively overcome existing quantum clustering technology in the limitation of handling on the large-scale data ability by the new affinity degree function calculation formula of design, can directly cut apart medical image;

3) the existing relatively quantum clustering technology of the present invention by proposing cloud mutation operation and hypercube interlace operation, can improve the segmentation precision of medical image, and diseased region is accurately located;

Simulation result shows that the present invention can effectively be applied to medical image to be cut apart, and human tissue structure and diseased region are accurately located, and helps improving the medical imaging diagnosis level.

Description of drawings

Fig. 1 is the medical image segmenting system synoptic diagram based on many elite's immunity quantum clusterings of the present invention;

Fig. 2 is the medical image dividing method realization flow figure based on many elite's immunity quantum clusterings of the present invention;

Fig. 3 is a cloud mutation operation synoptic diagram of the present invention;

Fig. 4 is a hypercube interlace operation synoptic diagram of the present invention;

Fig. 5 is that normal medical image of the present invention is cut apart simulation result;

Fig. 6 is that pathology medical image of the present invention is cut apart simulation result.

Embodiment

With reference to Fig. 1, the medical image segmenting system based on many elite's immunity quantum clusterings of the present invention mainly comprises: medical image pretreatment module, medical image data preparation module, medical image data cluster module and medical image segmentation result diagnostic module.Wherein:

The medical image pretreatment module, convert the medical image of rgb format to gray level image, and carry out the histogram equalization enhancement process, to the gradation of image value after handling according to pixels put from top to bottom, from left to right order forms a line and is input to the medical image data preparation module;

The medical image data preparation module is carried out cluster centre coding to the gray value data sample of input, forms the antibody of many elite's immunity quantum clusterings, and picked at random k antibody is as the initial antibodies population, k be medical image cut apart the classification number;

Medical image data cluster module, use many elite's immunity quantum clustering technical antagonism body populations and carry out cluster, and exporting cluster result to medical image segmentation result output module, it comprises: antibody affinity degree function design submodule, the sub-population of many elite are divided submodule and many elite immune optimization operation operator design submodule.This antibody affinity degree function design submodule, the schrodinger equation that is used for finding the solution quantum mechanics obtains the potential-energy function computing formula, and the potential-energy function minimal point here is corresponding with cluster centre, according to cluster centre calculating affinity degree functional value; The sub-population of these many elite is divided submodule, according to affinity degree ranking results antibody population is divided into sub-population of elite and common sub-population; This many elite immune optimization operation operator design submodule is used for carrying out successively the antibody cloning operation, the sub-population of elite is adopted the cloud mutation operation, disturbs the reorganization operation and utilize the hypercube interlace operation that sub-population of elite and common sub-population are carried out coevolution common sub-population employing is absolutely dry;

Medical image segmentation result output module turns back to the cluster result label of importing in the gray level image, exports final medical image segmentation result.

With reference to Fig. 2, the medical image dividing method based on many elite's immunity quantum clusterings of the present invention, concrete implementation step is as follows:

Step 1. pair medical image to be split carries out pre-service.

For the medical image of rgb format to be split, convert thereof into gray level image earlier.Because the medical image imaging results is subjected to the restriction of multiple factor to cause image blurring, human tissue organ's contrast and differentiates rate variance, therefore we strengthen medical image degree of comparing with the histogram equalization method again, again will be according to pixel from top to bottom through above pretreated gradation of image Value Data, order from left to right forms a line.

Step 2. pair medical image data carries out antibody coding and antibody population initialization.

Above-mentioned gray value data is carried out the cluster centre coding, form the antibody of many elite's immunity quantum clusterings, each antibody is a string real number, for the N dimension space, and k cluster centre, antibody length is N * k.Wherein, first length is that the field of N is represented first cluster centre, and the field that second length is N is represented second cluster centre, and the rest may be inferred.For example, consider 2 dimension data collection with 3 cluster centres, the length of antibody is exactly 2 * 3, and 3 sample points of picked at random are as initial center, for example, (1,3), (2,4) and (5,6), antibody just is encoded into 1-3-2-4-5-6, and each group is respectively represented an initial cluster center.A picked at random k antibody is as the initial antibodies population from the antibody behind the coding, k be medical image cut apart the classification number.

Step 3. is calculated the antibody affinity degree of many elite immune optimization.

(3.1) two antibody sample point x of definition _iAnd x _jBetween distance function D _Ij=|| x _i-x _j||, find the solution schrodinger equation

The computing formula that obtains potential-energy function V is:

$V=E-\frac{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">d</figure-callout>}}{2}+\frac{1}{{2\mathrm{\&sigma;}}^2}\frac{{\mathrm{\&Sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="j\; D\; ij"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">j</figure-callout>}}{D}_{\mathrm{ij}}^{}{}_2^{}\exp (-\frac{{\mathrm{<figure-callout\; id="2"\; label="D\; ij"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">D</figure-callout>}}_{\mathrm{ij}}^{}}{{2\mathrm{\&sigma;}}^2})}{{\mathrm{\&Sigma;}}_j\exp (-\frac{{\mathrm{<figure-callout\; id="2"\; label="D\; ij"\; filenames="H2009102186574E0000031.png"\; state="\{\{state\}\}">D</figure-callout>}}_{\mathrm{ij}}^{}}{{2\mathrm{\&sigma;}}^2})}$

Wherein, H is a Hamiltonian,

E is the energy feature value of Hamiltonian, and d is the intrinsic dimensionality of input data, and x is the input sample of data point, and σ is a scale parameter,

Be Laplace operator;

(3.2) determine K cluster centre { c according to the minimal point of the potential-energy function value that calculates _i, i=1 ... K}, and according to sample point and all kinds of division set Q of the nearest principle acquisition of each cluster centre Euclidean distance _i, then affinity degree function calculation formula is:

$f=1/(1+\underset{i=1}{\overset{K}{\mathrm{\&Sigma;}}}\underset{x_j\&Element;O_i}{\mathrm{\&Sigma;}}||x_j-c_i|\left|\right)$

Wherein, || || for asking for the operational symbol of Euclidean distance.Above-mentioned affinity degree function calculation formula only need calculate on the distance matrix basis between gray value data point, just can effectively overcome existing quantum clustering technology in the limitation of handling on the large-scale data ability.

The operation operator of many elite of step 4. design immune optimization.

(4.1) clone operations

Calculate the affinity degree value of all antibody in the antibody population and ordering from high to low, and higher pM the antibody of the back affinity degree value that will sort is as the sub-population of elite, remainder is as common sub-population, and wherein, p is the ratio value of setting, and M is the population size.Here " many elite " has two layers of meaning: the ground floor implication is meant that sub-population of elite and common sub-population all will keep the higher outstanding antibody of affinity degree value, and second layer implication is meant that the outstanding antibody of reservation is a plurality of rather than single.

To sub-population of elite and common sub-population, respectively according to the clone's scale N that sets _cA plurality of outstanding antibody are carried out clone operations.

(4.2) mutation operation

Cloud model is the qualitativing concept represented with natural language and the uncertain transformation model between its quantificational expression, mainly reflects the ambiguity and the randomness of notion in objective world or the human cognitive.Wherein, normal cloud model has randomness and steady tendency, the cloud model of the followed normal distribution stochastic distribution rule that characterizes with expectation value Ex, entropy En and super entropy He.Apply it in the mutation operation, make the randomness of cloud variation can keep the population diversity to avoid search to be absorbed in local extremum, steady tendency has then been protected defect individual preferably and has been carried out overall situation location.As shown in Figure 3, for one dimension normal cloud model C (0, En, He), be that Fig. 3 (a) is C (0 near 0 the true origin in expectation value, 0.5,0.1) time normal cloud model, Fig. 3 (b) is C (0,0.5,0.2) time normal cloud model, Fig. 3 (c) is C (0,1,0.1) normal cloud model the time, Fig. 3 (d) is C (0,1,0.2) normal cloud model the time.As seen, En is big more, and the water dust coverage is big more, and He is big more, and the water dust dispersion degree is big more.Expectation value has embodied the stability of cloud variation, and entropy has embodied the range of cloud variation, and super entropy has embodied the precision of cloud variation.

Adopt the concrete steps of cloud mutation operation as follows to the sub-population of elite:

At first calculate initial degree of certainty: μ=P according to affinity degree value _Max-(P _Max-P _Min) (f _Max-f)/(f _Max+ f _Min), wherein, f, f _MaxAnd f _MinBe respectively affinity degree value, affinity degree maximal value and the affinity degree minimum value of each iteration, P _MaxAnd P _MinBe respectively the degree of certainty maximal value and the degree of certainty minimum value of setting;

Then, make Ex be the variation before antibody, En=0.1 σ, σ are the standard deviation of each dimension data variable, He=0.1En;

Then, produce expectation and be En, variance is the normal distribution random number of He: En '=En+Herandn;

At last, between (0,1), produce a random number r, if satisfy r＜P _m, P _mBe the variation probability of setting, then the antibody after the variation is

(4.3) reorganization operation

For common sub-population, adopt based on quantum coherent characteristic structure absolutely dry and disturb the reorganization operation.Its advantage is when two antibody are the same, can avoid common reorganization to operate inoperative drawback, thereby make full use of antibody information as much as possible in the population.

General reorganization performance constraint is between two antibody, and when two antibody that participate in reorganization were identical, they just no longer proved effective.Disturb in the reorganization absolutely dry, all antibody all participate in reorganization in the population.For the antibody number is 5, and antibody length is 8 situation, and absolutely dry disturbing recombinated operation as shown in Table 1 and Table 2.Table 1 is the preceding antibody of reorganization, and table 2 is the antibody after recombinating.As seen, absolutely dry disturb the reorganization be a kind of recombination form that rearranges combination according to diagonal line, it can increase information interchange between antibody as much as possible, avoid the evolving precocity in later stage, overcome the locality and the one-sidedness of common reorganization operation, be applied in the image segmentation, to improve the global segmentation performance of medical image.

Table 1

Table 2

(4.4) selection operation

From forming new population through the outstanding antibody of selection the sub-population of antibody after clone operations, mutation operation and the reorganization operation.For the optimum antibody in each sub-population:

b _i(k)＝{a _ij(k)|f(a _ij(k))＝maxf(A _m(k))，j＝1，2，...，q _i}

b _i(k) replace initial antibodies A _i(k) probability is:

Wherein, β＞0 is a parameter relevant with the antibody population diversity, is worth greatly more, and diversity is good more.

(4.5) hypercube interlace operation

Through behind the selection operation, sub-population of elite and common sub-population affinity degree mxm. are respectively fitb1 and fitb2, when fitb1 〉=fitb2, the sub-population antibody of elite are carried out the hypercube intersection with pM minimum antibody of common sub-population affinity degree; When fitb1＜fitb2, pM the antibody that common sub-population affinity degree is the highest carries out hypercube with the sub-population antibody of elite and intersects.Shown in Fig. 4 (a), under the one-dimensional case, the search volume that hypercube intersects outwards is extended to line segment CD with line segment AB; Shown in Fig. 4 (b), under the two-dimensional case, the search volume is the expansion to the plane; Shown in Fig. 4 (c), under the three-dimensional situation, the search volume is made of jointly inner rectangular parallelepiped and outside continuation space; Three-dimensional above situation, the search volume then is to the continuation of hypercube according to corresponding dimension.

For intersecting parent antibody x _kAnd y _k, make l _Min=min (x _k, y _k), l _Max=max (x _k, y _k), δ=l _Max-l _Min, then be respectively x through the filial generation antibody after the hypercube interlace operation _K+1And y _K+1, wherein:

$x_{k+1}=\left\{\begin{array}{cc}\mathrm{unifrnd}(l_{\min }-\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;},l_{\max }+\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;})& r<P_c\\ x_k& \mathrm{else}\end{array}\right.$

$y_{k+1}=\left\{\begin{array}{cc}\mathrm{unifrnd}(l_{\min }-\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;},l_{\max }+\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;})& r<P_c\\ y_k& \mathrm{else}\end{array}\right.$

Wherein, r is the random number between (0,1), P _cBe crossover probability, unifrnd () is even value at random within the specific limits, and α=0.2 is the space continuation coefficient of hypercube.

After the hypercube interlace operation, return the clone operations of step (4.1) and carry out iteration optimization again, repeat N _MaxInferior.

Step 5. output medical image segmentation result.

Will be through N _MaxThe affinity degree mxm. fitb2 of affinity degree mxm. fitb1 of the sub-population of elite that inferior iteration obtains at last and common sub-population compares again, with the final cluster centre of the pairing antibody representative of the higher value after the comparison, and be divided in the different classifications according to this cluster centre each pixel with medical image, obtain final segmentation result.

Effect of the present invention can further specify by following emulation experiment figure:

Fig. 5 (a) is a width of cloth normal brain profile Magnetic resonance imaging.

Fig. 5 (b) uses the result that method of the present invention is cut apart to Fig. 5 (a).

Fig. 5 (c) is the normal eyeball blood vessel image of a width of cloth.

Fig. 5 (d) uses the result that method of the present invention is cut apart to Fig. 5 (c).

Fig. 6 (a) is the CT image of the adenocarcinoma of colon hepatic metastases of a width of cloth pathology.

Fig. 6 (b) uses the result that method of the present invention is cut apart to Fig. 6 (a).

White matter of brain among Fig. 5 (a), ectocinerea and circuitous around the profile that returns of ditch among Fig. 5 (b), come by division subtly.Particularly the fine structure at the cerebellum position in the corpus callosum in the middle of the figure and the lower right corner has obtained accurate location.

The human eye vascular distribution of Fig. 5 (c) is tight and meticulous, bifurcated uneven and and background difference little, from the segmentation result of Fig. 5 (d) as can be seen, method of the present invention is having good performance aspect the local detail processing.

Fig. 6 (a) is the CT image of a width of cloth adenocarcinoma of colon hepatic metastases, can know and see that belly transversal section CT scan image has shown a plurality of lumps, causes liver volume obviously to increase, and almost extends to left side edge from the human body right side.The centre of figure is the backbone and the sustainer of human body, can see the spleen of normal size from the lower right corner.As can be seen, use dividing method of the present invention from Fig. 6 (b), the diseased region of liver position high-density region is accurately detected, and, with the peripheral organs tissue good contrast is arranged.

Claims (4)

1. medical image segmenting system based on many elite immunity quantum clusterings comprises:

The medical image pretreatment module, convert the medical image of rgb format to gray level image, and carry out the histogram equalization enhancement process, and the gradation of image value after handling according to pixels to be put from top to bottom, order from left to right forms a line and is input to the medical image data preparation module;

The medical image data preparation module is carried out cluster centre coding to the gray value data sample of input, forms the antibody of many elite's immunity quantum clusterings, and picked at random k antibody is as the initial antibodies population, k be medical image cut apart the classification number;

Medical image data cluster module is used many elite's immunity quantum clustering antagonist populations and is carried out cluster, and exports cluster result to medical image segmentation result output module;

Medical image segmentation result output module turns back to the cluster result label of importing in the gray level image, exports final medical image segmentation result.

2. according to claims 1 described system, wherein medical image data cluster module comprises:

Antibody affinity degree function design submodule, the schrodinger equation that is used for finding the solution quantum mechanics obtains the potential-energy function computing formula, and this potential-energy function minimal point is corresponding with cluster centre, according to cluster centre calculating affinity degree functional value;

The sub-population of many elite is divided submodule, according to affinity degree functional value ranking results, antibody population is divided into sub-population of elite and common sub-population;

Many elite immune optimization operation operator design submodule is used for carrying out successively the antibody cloning operation, the sub-population of elite is adopted the cloud mutation operation, disturbs the reorganization operation and utilize the hypercube interlace operation that sub-population of elite and common sub-population are carried out coevolution common sub-population employing is absolutely dry.

3. the medical image dividing method based on many elite's immunity quantum clusterings comprises the steps:

(1) convert medical image to be split to gray level image, utilize histogram equalization method degree of comparing enhancement process, the gradation of image value after the enhancement process is according to pixels put from top to bottom, order from left to right forms a line;

(2) above-mentioned gray value data is carried out the cluster centre coding, form the antibody of many elite's immunity quantum clusterings, each antibody length is N * k, wherein, first length is that the field of N is represented first cluster centre, and the field that second length is N is represented second cluster centre, and the rest may be inferred;

(3) from the antibody behind the coding picked at random k antibody as the initial antibodies population, k be medical image cut apart the classification number;

(4) the affinity value of calculating antibody population:

4a) two antibody sample point x of definition _iAnd x _jBetween distance function D _Ij=|| x _i-x _j||, find the solution schrodinger equation The computing formula that obtains potential-energy function V is:

$V=E-\frac{d}{2}+\frac{1}{2{\mathrm{\&sigma;}}^2}\frac{{\mathrm{\&Sigma;}}_j{D}_{\mathrm{ij}}^2\exp (-\frac{D_{\mathrm{ij}}^2}{{2\mathrm{\&sigma;}}^2})}{{\mathrm{\&Sigma;}}_j\exp (-\frac{D_{\mathrm{ij}}^2}{{2\mathrm{\&sigma;}}^2})}$

Wherein, H is a Hamiltonian, E is the energy feature value of Hamiltonian, and d is the intrinsic dimensionality of input data, and x is the input sample of data point, and σ is a scale parameter,

Be Laplace operator;

4b) determine K cluster centre { c according to the minimal point of the potential-energy function value that calculates _i, i=1 ... K}, and according to sample point and all kinds of division set O of the nearest principle acquisition of each cluster centre Euclidean distance _i, then affinity degree function calculation formula is:

$f=1/(1+\underset{i=1}{\overset{K}{\mathrm{\&Sigma;}}}\underset{x_j\&Element;O_i}{\mathrm{\&Sigma;}}||x_j-c_i|\left|\right)$

Wherein, || || for asking for the operational symbol of Euclidean distance;

(5) calculate the affinity degree value of all antibody in the antibody population after, they are sorted from high to low, pM antibody getting the front is as the sub-population of elite, remainder is as common sub-population, wherein, the ratio value of p for setting, M is the population size;

(6) to the sub-population of elite and common sub-population according to the clone's scale N that sets _cCarry out clone operations;

(7) after the clone operations, the sub-population of elite is carried out the cloud mutation operation, common sub-population is carried out the absolutely dry reorganization operation of disturbing, concrete steps are as follows:

7a) produce several r between one (0,1) at random, if satisfy r＜P _m, P _mBe the variation probability of setting, then the sub-population antibody of elite after the variation is

Wherein, Ex is the preceding antibody of variation, and En '=En+Herandn, En=0.1 σ, σ are the standard deviation of each dimension data variable, and He=0.1En, randn represent to satisfy the random number of standardized normal distribution, μ=P _Max-(P _Max-p _Min) (f _Max-f)/(f _Max+ f _Min) expression degree of certainty value, wherein, f, f _MaxAnd f _MinBe respectively affinity degree value, affinity degree maximal value and the affinity degree minimum value of each iteration, p _MaxAnd p _MinBe respectively the degree of certainty maximal value and the degree of certainty value minimum value of setting;

7b) for common sub-population, adopt based on quantum coherent characteristic absolutely dry and disturb the reorganization operation, all antibody all participate in reorganization, carry out permutation and combination according to the diagonal line rule;

(8) the sub-population of antibody after operating through clone operations, mutation operation and reorganization, select outstanding antibody to form sub-population of new elite and common sub-population according to probability, the affinity degree mxm. of the sub-population of elite that this is new is fitb1, and the affinity degree mxm. of new common sub-population is fitb2;

(9) the affinity degree mxm. with the affinity degree mxm. of the new sub-population of elite and new common sub-population compares, if fitb1 〉=fitb2, pM just that the antibody of the sub-population of elite and common sub-population affinity degree is minimum antibody carries out hypercube and intersects; If fitb1＜fitb2, pM just that common sub-population affinity degree is the highest antibody carries out hypercube with the sub-population antibody of elite and intersects;

(10) through after the interlace operation, return step (4) and carry out iteration optimization again, repeat N _MaxInferior;

(11) will be through N _MaxThe affinity degree mxm. fitb2 of affinity degree mxm. fitb1 of the sub-population of elite that inferior iteration obtains at last and common sub-population compares again, represent final cluster centre with the pairing antibody of the higher value after the comparison, and be divided in the different classifications according to this cluster centre each pixel with medical image, obtain final segmentation result.

4. according to claims 3 described methods, the wherein described hypercube interlace operation of step (9), carry out according to following steps:

For two parent antibody x that intersect _kAnd y _k, make l _Min=min (x _k, y _k), l _Max=max (x _k, y _k), δ=l _Max-l _Min, then be respectively x through the filial generation antibody after the hypercube interlace operation _K+1And y _K+1, wherein:

$x_{k+1}=\left\{\begin{array}{cc}\mathrm{unifrnd}(l_{\min }-\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;},l_{\max }+\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;})& r<P_c\\ x_k& \mathrm{else}\end{array}\right.$

$y_{k+1}=\left\{\begin{array}{cc}\mathrm{unifrnd}(l_{\min }-\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;},l_{\max }+\mathrm{\&alpha;}\&CenterDot;\mathrm{\&delta;})& r<P_c\\ y_k& \mathrm{else}\end{array}\right.$

Wherein, r is the random number between (0,1), P _cBe crossover probability, the even at random within the specific limits value of unifrnd () expression, α=0.2 is the space continuation coefficient of hypercube.

Images