CN106097283A - A kind of multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic - Google Patents

Abstract

The invention discloses a kind of multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, including: gather digital X-ray image, and be normalized；Carry out gamma correction, color space conversion, obtain the image of XYZ color space, and carry out the simulation of hard copy；Gray proces, carries out channel decomposition, obtains cone image and retinal rod image；Laplacian pyramid algorithm is utilized respectively image to be carried out multi-resolution decomposition；Carry out brightness and contrast adaptive；Carry out sensing and threshold processing；Image merges respectively, and is combined rear image and carries out adverse transference sense and inverse gain process, it is achieved the image reconstruction of laplacian pyramid, obtains rebuilding image；Processing rebuilding image normalization, and normalization gained image is carried out color space conversion and gamma correction, rgb color space image after being corrected also exports.The present invention can carry out contrast enhancing effectively to X-ray image and details strengthens, it is possible to more preferably weakens the noise impact on X-ray image.

Description

A kind of multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic

Technical field

The present invention relates to a kind of multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, belong at image The technical field of reason.

Background technology

Breast carcinoma is the modal a kind of malignant tumor of women.Adding up according to World Health Organization (WHO), the whole world has more than 50 every year Ten thousand women die from breast carcinoma, are diagnosed with breast carcinoma more than 1,200,000 women, and sickness rate constantly rises, and increase crowd master If from developing country.Early detection is the most important means reducing Death Rate of Breast Cancer.Clinical screening and Diagnosis of Breast The method of cancer mainly have physical examination, serum female/progesterone receptor level measure and imaging diagnosis, wherein imaging diagnosis exists The inspection of breast carcinoma occupies very important status.And with X-ray photographic diagnosis of technique breast lesion, it is acknowledged as diagnosis breast The method most effective, most reliable of adenocarcinoma.

And during the radiophotography of X-ray image, owing to the x-ray in the complexity of breast tissue and imaging system dissipates Penetrate, the impact of the various unfavorable factors such as electro instrument noise, cause the decline of picture quality, mainly show as that details is fuzzy, contrast Difference, thus affect analysis and the diagnosis of doctor.Especially in breast carcinoma in early days, the sign Microcalcification granule of canceration is at x-ray figure Show the trickleest in Xiang, it is difficult to be noticeable.Therefore, X-ray image is carried out enhancement process, be doctor when diagnosing can not or The step lacked.

Traditional image enchancing method has histogram equalization process, edge enhancement process etc., and they can be effectively improved figure The contrast of picture, reaches to strengthen the visual effect of image.But, traditional image enchancing method can not strengthen x-ray figure effectively Details in Xiang, especially the best to the details reinforced effects of Microcalcification granule.Therefore the need that X-ray image details strengthens cannot be met Ask.

Summary of the invention

The technical problem to be solved is to overcome the deficiencies in the prior art, it is provided that a kind of based on human vision system The multiple dimensioned X-ray image Enhancement Method of system characteristic, can not strengthen in X-ray image solving traditional image enchancing method effectively Details, especially the best to the details reinforced effects of Microcalcification granule problem.

The present invention solves above-mentioned technical problem the most by the following technical solutions:

A kind of multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, including:

Step (1) gathers digital X-ray image A, and is normalized the pixel value of digital X-ray image A, To image

Step (2) is to described imageCarry out gamma correction, color space conversion, obtain the image of XYZ color space And by imagePixel be expressed as physical brightness value, obtain the analog image C of hard copy；

Step (3) carries out gray proces to the analog image C of described hard copy, according to the human eye cone, the spy of rod cell Property carries out channel decomposition, obtains cone image C_coneWith retinal rod image C_rod；

Step (4) utilizes laplacian pyramid algorithm respectively to cone image C_coneWith retinal rod image C_rodCarry out many chis Degree decomposes；

Step (5) is to the cone image C after described decomposition_coneWith retinal rod image C_rodCarry out brightness and contrast adaptive；

Step (6) is to the cone image C after described adaptation_coneWith retinal rod image C_rodCarry out sensing and threshold processing；

Cone image after step (6) processes and retinal rod image are merged by step (7) respectively, and are combined rear image and enter Row adverse transference sense and inverse gain process, it is achieved the image reconstruction of laplacian pyramid, obtain rebuilding image C '；

Described reconstruction image C ' is normalized by step (8), obtains the image after normalizationAnd to image Carry out color space conversion and gamma correction, the rgb color space image after being correctedAnd export.

Further, as a preferred technical solution of the present invention: the analog image C of hard copy in described step (2) Pixel value be:

$X=\stackrel{\·}{X}(L_{\max }-L_{\min })+X_W\·L_{\min }+L_{ambient}$

$Y=\stackrel{\‾}{Y}\·(L_{\max }-L_{\min })+Y_W\·L_{\min }+L_{ambient}$

$Z=\stackrel{\‾}{Z}\·(L_{\max }-L_{\min })+Z_W\·L_{\min }+L_{ambient}$

Wherein, (X, Y, Z) is the pixel value of image C；(X_W, Y_W, Z_W) it is the brightness value of white light；For image's Pixel value；L_maxMaximum for brightness；L_minMinima for brightness；L_ambientFor environmental light brightness.

Further, as a preferred technical solution of the present invention: cone image C in described step (3)_coneAnd retinal rod Image C_rodPixel value be:

$X=Y=Z=\frac{1}{3}(X+Y+Z)$

X_cone=0.3897X+0.6890Y-0.0787Z=X

X_rod=-0.702X+1.039Y+0.433Z=0.77X

Wherein, X_coneRepresent cone image C_conePixel value, X_rodRepresent retinal rod image C_rodPixel value.

Further, as a preferred technical solution of the present invention: described step (4) is by cone image C_coneCarry out many Scale Decomposition is:

LP_{Cone, n+1}=f_DOWN(LP_{Cone, n}), LP_{Cone, 0}=C_cone 0≤n≤N

BP_{Cone, n}=LP_{Cone, n}-f_UP(LP_{Cone, n+1}) 0≤n ＜ N

Wherein LP_{Cone, n}And BP_{Cone, n}Represent cone image C respectively_coneLow pass pyramid and the logical pyramidal n-th layer of band Image, N is pyramidal total number of plies, f_DOWNAnd f_UPRespectively down sample function and upwards interpolating function；

And, by retinal rod image C_rodCarrying out multi-resolution decomposition is:

LP_{Rod, n+1}=f_DOWN(LP_{Rod, n}), LP_{Rod, 0}=C_rod≤ n ＜ N

BP_{Rod, n}=LP_{Rod, n}-f_UP(LP_{Rod, n+1}) 0≤n ＜ N

Wherein LP_{Rod, n}And BP_{Rod, n}Represent retinal rod image C respectively_rodLow pass pyramid and the logical pyramidal n-th layer figure of band Picture, N is pyramidal total number of plies, f_DOWNAnd f_UPRespectively down sample function and upwards interpolating function.

Further, as a preferred technical solution of the present invention: described step (5) is to the cone figure after described decomposition As C_coneWith retinal rod image C_rodCarry out brightness and contrast adaptive, including,

Described band is led to pyramidal n-th layer image BP by step (5.1)_{Cone, n}And BP_{Rod, n}Calculate cone signal respectively and regard The gain factor G of bar signal_coneAnd G_rod；

Step (5.2) is by the gain factor G of tried to achieve cone signal_conePyramidal n-th layer image BP logical with band_{Cone, n} It is multiplied, obtains contrast adaptation image ABP_{Cone, n}；And the gain factor G by tried to achieve retinal rod signal_rodLogical with band pyramidal N-th layer image BP_{Rod, n}It is multiplied, obtains contrast adaptation image ABP_{Rod, n}；

Step (5.3) is according to the low pass pyramid n-th layer image LP of cone signal_{Cone, N}Calculate and obtain gain factor, and To this gain factor and low pass pyramidal n-th layer image LP_{Cone, N}Product；And the low pass pyramid N according to retinal rod signal Tomographic image LP_{Rod, N}Calculate and obtain gain factor, and obtain this gain factor and low pass pyramidal n-th layer image LP_{Rod, N}Take advantage of Long-pending, to complete adaptation.

Further, as a preferred technical solution of the present invention: described step (8) also includes according to output image Format adjusting imageValue.

Further, as a preferred technical solution of the present invention: described method strengthens for mammography X.

The present invention uses technique scheme, can produce following technique effect:

Based on human visual system's characteristic the multiple dimensioned X-ray image Enhancement Method that the present invention proposes, with existing method Compare, it is an advantage of the current invention that: (1) uses multiple dimensioned Enhancement Method can avoid the limitation of single yardstick Enhancement Method Property, grain effect will not be produced；(2) the multi-scale enhancement method based on human visual system's characteristic of the present invention, it is possible to for The sensitivity characteristic of human eye so that the reinforced effects of contrast and details is the best, is more suitable for eye-observation, it is also possible to more preferably Ground weakens the noise impact on X-ray image.

Therefore, the method for the present invention can carry out contrast enhancing effectively to X-ray image and details strengthens, and enables in particular to Strengthen Microcalcification granule, be suitably applied the diagnosis of breast carcinoma of early stage, it is possible to the diagnosis efficiency being effectively improved medical personnel is same Time reduce misdiagnosis rate, be the medical image enhancement method of a kind of practicality.

Accompanying drawing explanation

Fig. 1 is based on human visual system's characteristic the multiple dimensioned X-ray image Enhancement Method flow chart of the present invention.

Fig. 2 (a) is original image.

Fig. 2 (b) is the enhancing image of traditional histogram equalization method.

Fig. 2 (c) is the enhancing image using image averaging strategy in the inventive method.

Fig. 2 (d) is the enhancing image using pixel Average Strategy in the inventive method.

Detailed description of the invention

Below in conjunction with Figure of description, embodiments of the present invention are described.

As it is shown in figure 1, the present invention devises a kind of multiple dimensioned X-ray image enhancing side based on human visual system's characteristic Method, the method is applicable to medical image enhancement method, specifically includes following steps:

Step (1) gathers digital X-ray medical image.

Utilize equipment to gather digitized X-ray image A, and the pixel value of A is normalized, obtain

Step (2) carries out gamma correction, color space conversion and the simulation of hard copy to image, specific as follows:

First, to described imageCarry out gamma correction, obtain imageThat is:

$v=\left\{\begin{array}{cc}V/12.92& v\≤0.04045\\ {\left(\right(V+0.055)/1.055)}^{\mathrm{\γ}}& v>0.04045\end{array}\right.v\∈\{r,g,b\},V\∈\{R,G,B\}---\left(1\right)$

Wherein, v represents the image pixel value after correction, belongs to non-linear rgb color space；V represents the front image of correction Pixel value, belongs to linear rgb color space, and the span of γ is between 1.8-2.6.

Secondly, to gained imageCarry out color space conversion, as a example by standard light source D65, obtain XYZ color space ImageThat is:

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{ccc}2.3706743& -0.9000405& -0.4706338\\ -0.5138850& 1.4253036& 0.0885814\\ 0.0052982& -0.0146949& 1.0093968\end{array}\right]\left[\begin{array}{c}r\\ g\\ b\end{array}\right]---\left(2\right)$

Wherein (r, g b) represent imagePixel value, (X, Y, Z) representPixel value.

Finally, the maximum L of given brightness_maxWith minima L_minAnd environmental light brightness L_ambient, unit is candela Every square metre, by imagePixel value be expressed as physical brightness value, obtain the analog image C of hard copy, it may be assumed that

$X=\stackrel{\‾}{X}\·(L_{\max }-L_{\min })+X_W\·L_{\min }+L_{ambient}---\left(3\right)$

$Y=\stackrel{\‾}{Y}\·(L_{\max }-L_{\min })+Y_W\·L_{\min }+L_{ambient}---\left(4\right)$

$Z=\stackrel{\‾}{Z}\·(L_{\max }-L_{\min })+Z_W\·L_{\min }+L_{ambient}---\left(5\right)$

Wherein, (X_W, Y_W, Z_W) it is the brightness value of white light,For imagePixel value, (X, Y, Z) is image C's Pixel value.

Step (3) carries out channel decomposition according to the characteristic of the human eye cone, rod cell, is specially as follows:

First, the analog image C of hard copy is carried out gray proces, according to the human eye cone, the different qualities of rod cell Carry out channel decomposition, obtain cone image C_comeWith retinal rod image C_rod, i.e.

$X=Y=Z\frac{1}{3}(X+Y+Z)---\left(6\right)$

X_cone=0.3897X+0.6890Y-0.0787Z=X (7)

X_rod=-0.702X+1.039Y+0.433Z=0.77X (8)

Wherein, X_coneRepresent cone image C_conePixel value, X_rodRepresent retinal rod image C_rodPixel value.

Step (4) utilizes laplacian pyramid algorithm respectively cone image and retinal rod figure to be carried out multi-resolution decomposition, tool Body is as follows:

First, utilize laplacian pyramid algorithm to cone image C_coneCarry out multi-resolution decomposition, it may be assumed that

LP_{Cone, n+1}=f_DOWN(LP_{Cone, n}), LP_{Cone, 0}=C_cone0≤n ＜ N (9)

BP_{Cone, n}=LP_{Cone, n}-f_UP(LP_{Cone, n+1}) 0≤n≤N (10)

Wherein, LP_{Cine, n}And BP_{Cone, n}Represent cone image C respectively_coneLow pass pyramid and band logical pyramidal n-th Tomographic image, N is pyramidal total number of plies, f_DOWNAnd f_UPRespectively down sample function and upwards interpolating function.

In like manner, to retinal rod image C_rodCarry out multi-resolution decomposition, it may be assumed that

LP_{Rod, n+1}=f_DOWN(LP_{Rod, n}), LP_{Rod, 0}=C_rod0≤n ＜ N (11)

BP_{Rod, n}=LP_{Rod, n}-f_UP(LP_{Rod, n+1}) 0≤n ＜ N (12)

Wherein, LP_{Rod, n}And BP_{Rod, n}Represent retinal rod image C respectively_rodLow pass pyramid and the logical pyramidal n-th layer of band Image, N is pyramidal total number of plies, f_DOWNAnd f_UPRespectively down sample function and upwards interpolating function.

It is adaptive that step (5) carries out brightness and contrast to the cone image after described decomposition and retinal rod image, specific as follows:

Firstly, for cone image C_coneBand lead to pyramidal n-th layer image BP_{Cone, n}With retinal rod image C_rodThe logical gold of band The n-th layer image BP of word tower_{Rod, n}, calculate the gain factor G of cone signal and retinal rod signal respectively_coneAnd G_rodCalculating public Formula, it may be assumed that

$G_{cone}\left(I\right)=\frac{1}{0.555{(I+1.0)}^{0.85}}---\left(13\right)$

$G_{rod}\left(I\right)=\[\frac{10}{I^2+10}\]\[\frac{1}{0.908{(I+0.001)}^{0.85}}\]---\left(14\right)$

Wherein, I is input signal.

Image BP logical for band_{Cone, n}And BP_{Rod, n}, by corresponding low-pass pictures LP_{Cone, n+1}And LP_{Rod, n+1}As input letter Number substitute into formula (13) and (14) respectively, the gain factor G tried to achieve_cone(LP_{Cone, n+1}) and G_rod(LP_{Rod, n+1}) respectively with BP_{Cone, n}And BP_{Rod, n}It is multiplied, obtains contrast adaptation image ABP_{Cone, n}And ABP_{Rod, n}, it may be assumed that

ABP_{Cone, n}=G_cone(LP_{Cone, n+1})·BP_{Cone, n}, 0≤n ＜ N (15)

ABP_{Rod, n}=G_rod(LP_{Rod, n+1})·BP_{Rod, n}, 0≤n ＜ N (16)

And for low-pass pictures LP_{Cone, N}And LP_{Rod, N}, it is divided into two kinds of gain control strategies: the first image averaging strategy, First try to achieve the meansigma methods of overall brightness in low-pass picturesWithObtain corresponding gain factor againWithThe final product that adaptive result is gain factor and low-pass pictures, it may be assumed that

${\mathrm{ALP}}_{cone}=G_{cone}\left(\stackrel{\‾}{{\mathrm{LP}}_{cone,N}}\right)\·{\mathrm{LP}}_{cone,N}---\left(17\right)$

${\mathrm{ALP}}_{rod}=G_{rod}\left(\stackrel{\‾}{{\mathrm{LP}}_{rod,N}}\right)\·{\mathrm{LP}}_{rod,N}---\left(18\right)$

The second pixel Average Strategy, respectively to low-pass pictures LP_{Cone, N}And LP_{Rod, N}Each pixel carry out gain factor Calculate, obtain corresponding gain factor figure, the final product that adaptive result is gain factor figure and low-pass pictures, it may be assumed that

ALP_cone=G_cone(LP_{Cine, n})·LP_{Cine, N} (19)

ALP_rod=G_rod(LP_{Rod, N})·LP_{Rod, N} (20)

Cone image after described adaptation and retinal rod image are sensed and threshold processing by step (6), specific as follows:

Firstly, for the logical image of band, cone signal and output valve T of retinal rod sensing_{BP, cone, n}And T_{BP, rod, n}Respectively For:

Wherein, 0≤n ＜ N, the value of index p is as shown in the table with the corresponding relation of number of plies n:

Then, carry out threshold processing, i.e. work as T_{BP, cone, n}Or T_{BP, rod, n}Value more than 50 time, it is set to 50.

Then, for low-pass pictures TLP_coneAnd TLP_rod, the most only need to calculate sensing defeated of cone signal and retinal rod signal Go out to be worth T_{LP, cone}And T_{LP, rod}, it is not necessary to threshold processing, it may be assumed that

T_{LP, cone}=ALP_cone/|ALP_cone|×30.5(ALP_cone)^0.5 (23)

T_{LP, rod}=ALP_rod/|ALP_rod|×122(ALP_rod)^0.5 (24)

Step 7: image is merged respectively, and be combined rear image and carry out adverse transference sense process and inverse gain process, real The image reconstruction of existing laplacian pyramid, obtains rebuilding image.Specific as follows:

First, by corresponding cone image T_{BP, cone, n}And T_{LP, cone}And retinal rod image T_{BP, rod, n}And T_{LP, rod}Close respectively And, it may be assumed that

T_{BP, n}=T_{BP, cone, n}+T_{BP, rod, n}/ 7,0≤n ＜ N (25)

T_LP=T_{LP, cone}+T_{LP, rod}/7 (26)

Wherein, T_{BP, n}And T_LPIt is respectively the band after merging and leads to image and low-pass pictures.

Then, the image after being combined carries out adverse transference sense process, it may be assumed that

${{\mathrm{ALP}}_N}^{\′}=T_{LP}/\left|T_{LP}\right|\×{\left(\frac{T_{LP}}{30.5}\right)}^2---\left(28\right)$

Wherein, ABP_n' and ALP_N' it is respectively the high fdrequency component of the contrast adaptation image after adverse transference sense processes and low frequency division Amount.

Finally, carry out inverse gain process and image reconstruction, provide inversion gain factor G_inv(I) formula, it may be assumed that

G_inv(I)=0.555 (I+1.0)^0.85 (29)

By the mean flow rate of display screenAs input signal, calculate the inverse gain factor of this signal, try to achieve ALP_N' and should The product LP of inverse gain factor_N', it may be assumed that

${\mathrm{LP}}_N^{\′}=G_{inv}\left(\stackrel{\‾}{D}\right)\·{{\mathrm{ALP}}_N}^{\′}=0.555{(D+1.0)}^{0.85}\·{\mathrm{ALP}}_N^{\′}---\left(30\right)$

By LP_N' as the first step rebuild, carry out the inverse gain process with logical image and image reconstruction, in the process The remaining image LP ' in low pass pyramid is tried to achieve successively according to formula (31)_N-1, LP '_N-2..., LP₀', finally give reconstruction figure As C ', it may be assumed that

LP_n'=G_inv(LP_n+1′)ABP_n′+LP_n+1', 0≤n≤N (31)

C '=LP₀′ (32)

Step 8: be normalized described reconstruction image, obtains the image after normalization, and image is carried out color Color space conversion and gamma correction, rgb color space image after being corrected also exports.Specific as follows:

First, the image C ' after rebuilding is normalized, obtains the image after normalizationTo imageCarry out Color space is changed, and obtains imageThat is:

$\left[\begin{array}{c}{r}^{\′}\\ g^{\′}\\ b^{\′}\end{array}\right]=\left[\begin{array}{ccc}2.3706743& -0.9000405& -0.4706338\\ -0.5138850& 1.4253036& 0.0885814\\ 0.0052982& -0.0146949& 1.0093968\end{array}\right]\left[\begin{array}{c}{X}^{\′}\\ Y^{\′}\\ Z^{\′}\end{array}\right]---\left(33\right)$

Wherein, (r ', g ', b ') representsPixel value, (X ', Y ', Z ') representPixel value.Then to imageEnter Row gamma correction, the image after being correctedThat is:

$V^{\′}=\left\{\begin{array}{cc}12.92v^{\′}& v^{\′}\≤0.0031308\\ 1.055v^{\′\frac{1}{\mathrm{\γ}}}-0.555& v^{\′}>0.0031308\end{array}\right.v^{\′}\∈\{r^{\′},g^{\′},b^{\′}\},V^{\′}\∈\{R^{\′},G^{\′},B^{\′}\}---\left(34\right)$

Wherein, v ' represents the image pixel value before correction, belongs to non-linear rgb color space, and V ' represents image after correction Pixel value, belong to linear rgb color space, the span of γ is between 1.8-2.6.

Finally, it is preferable that according to the Format adjusting image of output imageValue, export image A ', it may be assumed that

$A^{\′}=(2^n-1)\·{\stackrel{\‾}{A}}^{\′}---\left(35\right)$

Wherein, n is the figure place of image.

It addition, the present embodiment gives the simulation comparison example before and after X-ray image strengthens, as shown in Fig. 2 (a), for original Image；As shown in Fig. 2 (b), for the enhancing image of traditional histogram equalization method；As shown in Fig. 2 (c), for side of the present invention Method uses the enhancing image of image averaging strategy；As shown in Fig. 2 (d), for the inventive method uses pixel Average Strategy Strengthen image.Thus, it is possible to find out Fig. 2 (c) and Fig. 2 (d) so that the reinforced effects of picture contrast and details is the best Good, it is possible to preferably to weaken the noise impact on X-ray image.

Thus, based on human visual system's characteristic the multiple dimensioned X-ray image Enhancement Method that the present invention proposes can be effective Ground carries out contrast enhancing to X-ray image and details strengthens, and enables in particular to strengthen Microcalcification granule, is suitably applied mammary gland in early days The diagnosis of cancer, is effectively used in mammography X enhancing.The diagnosis efficiency that can be effectively improved medical personnel is same Time reduce misdiagnosis rate, be the medical image enhancement method of a kind of practicality.

Above in conjunction with accompanying drawing, embodiments of the present invention are explained in detail, but the present invention is not limited to above-mentioned enforcement Mode, in the ken that those of ordinary skill in the art are possessed, it is also possible on the premise of without departing from present inventive concept Make a variety of changes.

Claims (7)

1. a multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, it is characterised in that including:

Step (1) gathers digital X-ray image A, and is normalized the pixel value of digital X-ray image A, obtains figure Picture

Step (2) is to described imageCarry out gamma correction, color space conversion, obtain the image of XYZ color spaceAnd will figure PicturePixel be expressed as physical brightness value, obtain the analog image C of hard copy；

Step (3) carries out gray proces to the analog image C of described hard copy, enters according to the characteristic of the human eye cone, rod cell Row of channels is decomposed, and obtains cone image C_coneWith retinal rod image C_rod；

Step (4) utilizes laplacian pyramid algorithm respectively to cone image C_coneWith retinal rod image C_rodCarry out multiple dimensioned point Solve；

Step (5) is to the cone image C after described decomposition_coneWith retinal rod image C_rodCarry out brightness and contrast adaptive；

Step (6) is to the cone image C after described adaptation_coneWith retinal rod image C_rodCarry out sensing and threshold processing；

Cone image after step (6) processes and retinal rod image are merged by step (7) respectively, and are combined rear image and carry out inverse Sensing and inverse gain process, it is achieved the image reconstruction of laplacian pyramid, obtain rebuilding image C '；

Described reconstruction image C ' is normalized by step (8), obtains the image after normalizationAnd to imageCarry out Color space conversion and gamma correction, the rgb color space image after being correctedAnd export.

The most according to claim 1, multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, its feature exists In: in described step (2), the pixel value of the analog image C of hard copy is:

$X=\stackrel{\‾}{X}\·(L_{max}-L_{\min })+X_W\·L_{\min }+L_{ambient}$

$Y=\stackrel{\‾}{Y}\·(L_{max}-L_{min})+Y_W\·L_{\min }+L_{ambient}$

$Z=\stackrel{\‾}{Z}\·(L_{max}-L_{min})+Z_W\·L_{\min }+L_{ambient}$

Wherein, (X, Y, Z) is the pixel value of image C；(X_W, Y_W, Z_W) it is the brightness value of white light；For imagePixel Value；L_maxMaximum for brightness；L_minMinima for brightness；L_ambientFor environmental light brightness.

The most according to claim 1, multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, its feature exists In: cone image C in described step (3)_coneWith retinal rod image C_rodPixel value be:

$X=Y=Z=\frac{1}{3}(X+Y+Z)$

X_cone=0.3897X+0.6890Y-0.0787Z=X

X_rod=-0.702X+1.039Y+0.433Z=0.77X

Wherein, X_coneRepresent cone image C_conePixel value, X_rodRepresent retinal rod image C_rodPixel value.

The most according to claim 1, multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, its feature exists In: described step (4) is by cone image C_coneCarrying out multi-resolution decomposition is:

LP_{Cone, n+1}=f_DOWN(LP_{Cone, n}), LP_{Cone, 0}=C_cone0≤n ＜ N

BP_cone,n=LP_cone,n-fup(LP_cone,n+1) 0≤n ＜ N

Wherein LP_{Cone, n}And BP_{Cone, n}Represent cone image C respectively_coneLow pass pyramid and the logical pyramidal n-th layer image of band, N is pyramidal total number of plies, f_DOWNAnd f_UPRespectively down sample function and upwards interpolating function；

And, by retinal rod image C_rodCarrying out multi-resolution decomposition is:

LP_{Rod, n+1}=f_DOWN(LP_{Rod, n}), LP_{Rod, 0}=C_rod0≤n ＜ N

BP_{Rod, n}=LP_{Rod, n}-f_UP(LP_{Rod, n+1}) 0≤n ＜ N

Wherein LP_{Rod, n}And BP_{Rod, n}Represent retinal rod image C respectively_rodLow pass pyramid and the logical pyramidal n-th layer image of band, N For pyramidal total number of plies, f_DOWNAnd f_UPRespectively down sample function and upwards interpolating function.

The most according to claim 4, multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, its feature exists In: described step (5) is to the cone image C after described decomposition_coneWith retinal rod image C_rodCarry out brightness and contrast adaptive, bag Include,

Described band is led to pyramidal n-th layer image BP by step (5.1)_{Cone, n}And BP_{Rod, n}Calculate cone signal and retinal rod letter respectively Number gain factor G_coneAnd G_rod；

Step (5.2) is by the gain factor G of tried to achieve cone signal_conePyramidal n-th layer image BP logical with band_{Cone, n}It is multiplied, Obtain contrast adaptation image ABP_{Cone, n}；And the gain factor G by tried to achieve retinal rod signal_rodPyramidal n-th layer logical with band Image BP_{Rod, n}It is multiplied, obtains contrast adaptation image ABP_{Rod, n}；

Step (5.3) is according to the low pass pyramid n-th layer image LP of cone signal_{Cone, N}Calculate and obtain gain factor, and be somebody's turn to do Gain factor and low pass pyramidal n-th layer image LP_{Cone, N}Product；And the low pass pyramid n-th layer figure according to retinal rod signal As LP_{Rod, N}Calculate and obtain gain factor, and obtain this gain factor and low pass pyramidal n-th layer image LP_{Rod, N}Product, To complete adaptation.

The most according to claim 1, multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, its feature exists In: described step (8) also includes the Format adjusting image according to output imageValue.

The most according to claim 1, multiple dimensioned X-ray image Enhancement Method based on human visual system's characteristic, its feature exists In: described method strengthens for mammography X.