CN104574361A - Image processing method and device for mammary peripheral tissue equalization - Google Patents

Image processing method and device for mammary peripheral tissue equalization Download PDF

Info

Publication number
CN104574361A
CN104574361A CN201410707789.4A CN201410707789A CN104574361A CN 104574361 A CN104574361 A CN 104574361A CN 201410707789 A CN201410707789 A CN 201410707789A CN 104574361 A CN104574361 A CN 104574361A
Authority
CN
China
Prior art keywords
image
high frequency
gain
frequency imaging
layer
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201410707789.4A
Other languages
Chinese (zh)
Other versions
CN104574361B (en
Inventor
李海春
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Neusoft Medical Systems Co Ltd
Original Assignee
Neusoft Medical Systems Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Neusoft Medical Systems Co Ltd filed Critical Neusoft Medical Systems Co Ltd
Priority to CN201410707789.4A priority Critical patent/CN104574361B/en
Publication of CN104574361A publication Critical patent/CN104574361A/en
Application granted granted Critical
Publication of CN104574361B publication Critical patent/CN104574361B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20048—Transform domain processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30068—Mammography; Breast

Abstract

The embodiment of the invention discloses an image processing method and device for mammary peripheral tissue equalization. The method comprises the following steps: extracting a mask image of effective human tissue from a mammary original image; carrying out grey level transformation on the original image so as to obtain an image to be processed; decomposing the image to be processed so as to obtain a low-frequency image and a plurality of high-frequency images; calculating according to the percentage composition of the mask image, the image to be processed and a gland so as to obtain a grey mapping curve; carrying out grey mapping on the low-frequency image through the grey mapping curve so as to obtain a mapped low-frequency image; carrying out gain processing on all high-frequency images so as to obtain gained high-frequency images; carrying out reconstruction processing on the mapped low-frequency image and the gained high-frequency images, and thus obtaining an equalized image. The method can carry out adaptive adjustment according to individual differences, thereby ensuring the equalization effect, reducing noises and artifacts, and improving the processing efficiency.

Description

The image processing method of a kind of mammary gland peripheral tissues equilibrium and device
Technical field
The present invention relates to field of medical image processing, be specifically related to image processing method and the device of the equilibrium of a kind of mammary gland peripheral tissues.
Background technology
Breast tissue characteristic distributions causes central area and fringe region to there is gray difference on a large scale, part tissue only can be observed according to fixing luminance contrast, according to dynamic range compression method, the little portfolio effect that easily causes of suppressed range is not enough, suppressed range is excessive can cause the problem that noise, artifact are excessively strong again, also may change the original gray distribution features in gland tissue region time serious.
In prior art, usual use large scale convolution kernel method is balanced to mammary gland peripheral tissues, the method specifically adopts fixing compressibility coefficient to make bulk treatment to image, due to mammary gland near breastwork side organizational information and peripheral tissues's density variation larger, adopt fixing compressibility coefficient, compressibility coefficient is excessive, body of gland region original gradation feature may change, the too small portfolio effect of compressibility coefficient is not enough, peripheral tissues can not be seen clearly completely, therefore, fixing compressibility coefficient cannot ensure that in image, all organizing all is processed well, in addition, large scale convolution kernel method often adopts convolution kernel size to reach 100 ~ 300, cause the calculated amount of this equilibrium treatment very large, treatment effeciency is lower.
Therefore, how to carry out rationally equilibrium treatment efficiently for mammary gland peripheral tissues and become the technical matters paid close attention to most in breast tissue observation clinical practice.
Summary of the invention
In order to solve the problems of the technologies described above, the invention provides image processing method and the device of a kind of mammary gland peripheral tissues equilibrium, not only ensure portfolio effect but also noise decrease and artifact, having also improved treatment effeciency.
In first aspect present invention, provide the image processing method of a kind of mammary gland peripheral tissues equilibrium, described method comprises:
From mammary gland original image, extract human body effectively organize mask image;
Greyscale transformation is carried out to described original image and obtains pending image, decompose described pending image and obtain a low-frequency image and multiple high frequency imaging;
Grey scale mapping curve is calculated according to described mask image, described pending image and body of gland percentage composition;
Adopt grey scale mapping curve to carry out grey scale mapping to low-frequency image to obtain mapping rear low-frequency image;
As gain process, high frequency imaging after gain is obtained to each high frequency imaging;
As reconstruction processing, the image after equilibrium is obtained to high frequency imaging after low-frequency image after described mapping and described gain.
Preferably, describedly calculate grey scale mapping curve according to mask image, pending image and body of gland percentage composition, comprising:
According to the effective tissue regions histogram of mask image statistics human body, according to described histogram, described body of gland percentage composition and preset peripheral tissues number percent definite threshold Tg and Ta, and according to body of gland percentage composition determination compressibility coefficient Kg and Ka; Wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
Determine at least three unique points according to Kg, Ka, Tg, Ta and described pending gradation of image maximal value maxP, and generate grey scale mapping curve according to described at least three unique point matchings.
Preferably, described as gain process, high frequency imaging after gain is obtained to each high frequency imaging, comprising:
The low-frequency image half-tone information corresponding according to each high frequency imaging determines the first gain coefficient that each high frequency imaging is corresponding;
The second gain coefficient that each high frequency imaging is corresponding is determined according to the half-tone information of each high frequency imaging;
According to formula S i'=Gain2i* (Gain1i*S i-S i)+S i, calculate high frequency imaging after gain; Wherein, Gain2i represents that the second gain coefficient that i-th layer of high frequency imaging is corresponding, Gain1i represent the first gain coefficient that i-th layer of high frequency imaging is corresponding, S irepresent i-th layer of high frequency imaging, S i' represent i-th layer of high frequency imaging after gain, the span of i is 1,2 ... N-1, N, N refers to the maximum number of plies that described multi-resolution pyramid algorithm adopts when decomposing pending image.
Preferably, the half-tone information of the described low-frequency image corresponding according to each high frequency imaging is determined to comprise the first gain coefficient that each high frequency imaging is corresponding:
The low-frequency image corresponding according to every layer of high frequency imaging determines corresponding Tg and Ta, and wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
According to formula F gi=(FgN-Fa) * i/N+Fa, calculate the corresponding parameter Fgi in i-th layer of high frequency imaging body of gland region, wherein FgN is the corresponding parameter in n-th layer high frequency imaging body of gland region, and FgN span is 1 ~ 5, Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1;
According to the i-th floor height image characteristic point A1 (0 frequently, Fgi), A2 (Tg, Fgi), A3 (Ta, and A4 (maxP Fa), Fa) matching obtains the first gain trace, determines according to described first gain trace the first gain coefficient that i-th layer of high frequency imaging is corresponding.
Preferably, the described half-tone information according to each high frequency imaging is determined to comprise the second gain coefficient that each high frequency imaging is corresponding:
Calculate human body and effectively organize interior high-frequency gain maximal value max (abs (Si (ROI))), mean value ts1 and two times mean value ts2, and determine attenuation coefficient Fs, this attenuation coefficient Fs span is 0.5 ~ 1, and Fs is greater than Fa; Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1, and wherein, Si (ROI) is that the human body of the i-th floor high frequency imaging effectively organizes district;
According to the i-th floor height image characteristic point B1 (0,1), B2 (ts1,1), B3 (ts2, Fs) and B4 (max (abs (S frequently i(ROI))), Fa) calculate the second gain trace, determine according to described second gain trace the second gain coefficient that i-th layer of high frequency imaging is corresponding.
Preferably, described as reconstruction processing, the image after equilibrium is obtained to high frequency imaging after low-frequency image after described mapping and described gain, comprising:
Realize the reconstruction processing of i-th layer in the following manner successively, wherein i successively value be N, N-1, N-2 ... 2,1:
The low-frequency image corresponding to i-th layer of high frequency imaging is made interpolation convolution and is obtained low-frequency image GR i, then calculate i-th layer of high frequency imaging S after described low-frequency image GRi gain i' between with value GS i; Every i layer reconstruction result is as low-frequency image corresponding to the i-th-1 layer high frequency imaging;
Carry out down-sampledly obtaining down-sampled mask image to mask image, determine region, image border Edge according to down-sampled mask image i;
According to GS i* Edge i> I bithe outline region of marking image fringe region, wherein I biit is signature grey scale value in the background of the low-frequency image that i-th layer of high frequency imaging is corresponding;
According to formula GP i=(GR i+ (k*S i' * Edge i+ S i' * (1-Edge i))), wherein k span is 0 ~ 1, calculates i-th layer of reconstruction result GP i, then by reconstruction result GP ithe pixel assignment of China and foreign countries' contour area is I bi, realize the reconstruction processing of i-th layer successively until complete the 1st layer of reconstruction processing obtain equilibrium after image.
Preferably, the described pending image of described decomposition obtains a low-frequency image and multiple high frequency imaging is specially: adopt multi-resolution pyramid algorithm to decompose described pending image and obtain a low-frequency image and multiple high frequency imaging.
In second aspect present invention, provide the image processing apparatus of a kind of mammary gland peripheral tissues equilibrium, described device comprises:
Extraction unit, effectively organizes mask image for extracting human body from mammary gland original image;
Resolving cell, obtains pending image for carrying out greyscale transformation to described original image, decomposes described pending image and obtains a low-frequency image and multiple high frequency imaging;
Grey scale mapping curve unit, for calculating grey scale mapping curve according to described mask image, described pending image and body of gland percentage composition;
Map unit, carries out grey scale mapping for adopting grey scale mapping curve to low-frequency image and obtains mapping rear low-frequency image;
Gain processing unit, for obtaining high frequency imaging after gain to each high frequency imaging as gain process;
Reconfigurable processing unit, for obtaining the image after equilibrium to high frequency imaging after low-frequency image after described mapping and described gain as reconstruction processing.
Preferably, described grey scale mapping curve unit, comprising:
Threshold value determination subelement, for adding up the effective tissue regions histogram of human body according to mask image, according to described histogram, described body of gland percentage composition and preset peripheral tissues number percent definite threshold Tg and Ta, and according to body of gland percentage composition determination compressibility coefficient Kg and Ka; Wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
Matching subelement, for determining at least three unique points according to Kg, Ka, Tg, Ta and described pending gradation of image maximal value maxP, and generates grey scale mapping curve according to described at least three unique point matchings.
Preferably, described gain processing unit, comprising:
First gain coefficient determination subelement, determines for the low-frequency image half-tone information corresponding according to each high frequency imaging the first gain coefficient that each high frequency imaging is corresponding;
Second gain coefficient determination subelement, for determining the second gain coefficient that each high frequency imaging is corresponding according to the half-tone information of each high frequency imaging;
Gain process subelement, for according to formula S i'=Gain2i* (Gain1i*S i-S i)+S i, calculate high frequency imaging after gain; Wherein, Gain2i represents that the second gain coefficient that i-th layer of high frequency imaging is corresponding, Gain1i represent the first gain coefficient that i-th layer of high frequency imaging is corresponding, S irepresent i-th layer of high frequency imaging, S i' represent i-th layer of high frequency imaging after gain, the span of i is 1,2 ... N-1, N, N refers to the maximum number of plies that described multi-resolution pyramid algorithm adopts when decomposing pending image.
Preferably, described first gain coefficient determining unit, comprising:
Intensity slicing value determination subelement, determine corresponding Tg and Ta for the low-frequency image corresponding according to every layer of high frequency imaging, wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
The corresponding parameter determination in region subelement, for according to formula F gi=(FgN-Fa) * i/N+Fa, calculate the corresponding parameter Fgi in i-th layer of high frequency imaging body of gland region, wherein FgN is the corresponding parameter in n-th layer high frequency imaging body of gland region, FgN span is 1 ~ 5, Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1;
First gain coefficient computation subunit, for according to the i-th floor height frequently image characteristic point A1 (0, Fgi), A2 (Tg, Fgi), A3 (Ta, and A4 (maxP Fa), Fa) calculate the first gain trace, determine according to described first gain trace the first gain coefficient that i-th layer of high frequency imaging is corresponding.
Preferably, described second gain coefficient determination subelement, comprising:
Parameter computation unit, interior high-frequency gain maximal value max (abs (Si (ROI))), mean value ts1 and two times mean value ts2 is effectively organized for calculating human body, and determine attenuation coefficient Fs, this attenuation coefficient span is 0.5 ~ 1, and Fs is greater than Fa; Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1;
Wherein, Si (ROI) is that the human body of the i-th floor high frequency imaging effectively organizes district;
Second gain coefficient computation subunit, for according to the i-th floor height frequently image characteristic point B1 (0,1), B2 (ts1,1), B3 (ts2, Fs) and B4 (max (abs (S i(ROI))), Fa) matching obtains the second gain trace, determines according to described second gain trace the second gain coefficient that i-th layer of high frequency imaging is corresponding.
Preferably, described reconfigurable processing unit, comprising:
With value computation subunit, make interpolation convolution for the low-frequency image corresponding to i-th layer of high frequency imaging and obtain low-frequency image GR i,calculate i-th layer of high frequency imaging S after described low-frequency image GRi and gain again i' between with value GS i, every i layer reconstruction result is as low-frequency image corresponding to the i-th-1 layer high frequency imaging; Wherein i successively value be N, N-1, N-2 ... 2,1;
Subelement is determined in region, image border, down-sampledly obtaining down-sampled mask image, determining region, image border Edge according to down-sampled mask image for carrying out mask image i;
Region, image border outline determination subelement, for according to GS i* Edge i> I bithe outline region of marking image fringe region, wherein I biit is signature grey scale value in the background of the low-frequency image that i-th layer of high frequency imaging is corresponding;
Reconstruction processing subelement, for according to formula GP i=(GR i+ (k*S i' * Edge i+ S i' * (1-Edge i))), wherein k span is 0 ~ 1, calculates i-th layer of reconstruction result GP i, then by reconstruction result GP ithe pixel assignment of China and foreign countries' contour area is I bi, realize the reconstruction processing of i-th layer successively until complete the 1st layer of reconstruction processing obtain equilibrium after image.
Technical scheme provided by the invention, compared with existing technical scheme, has following beneficial effect:
The body of gland composition that the present invention is based on image calculates grey scale mapping curve, utilizes this grey scale mapping curve to carry out mapping process to low-frequency image, like this, can reduce the gray difference of central area and peripheral tissues; Then again gain process is carried out to each layer of high frequency imaging, like this, make body of gland region and peripheral tissues's high-frequency information more balanced, finally, the image after obtaining equilibrium is reconstructed to the high frequency imaging after low-frequency image and gain process.Because grey scale mapping curve is based on the curve that obtains of body of gland composition matching, on the one hand, can under the prerequisite ensureing the distribution of body of gland primitive character, can the intensity of conservative control equilibrium; On the other hand, owing to considering the difference of image individuality, improve the consistance of result to a certain extent.In addition, because the present invention carries out multiple dimensioned decomposition to pending image, equilibrium treatment can be carried out under multiple yardstick, there is higher treatment effeciency.
Accompanying drawing explanation
In order to be illustrated more clearly in the technical scheme in the embodiment of the present application, below the accompanying drawing used required in describing embodiment is briefly described, apparently, accompanying drawing in the following describes is only some embodiments of the application, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the process flow diagram of the image processing method of the mammary gland peripheral tissues equilibrium of the embodiment of the present invention one;
Fig. 2 is the process flow diagram of the image processing method of the mammary gland peripheral tissues equilibrium of the embodiment of the present invention two;
Fig. 3 is the structural drawing of the image processing apparatus of the mammary gland peripheral tissues equilibrium of the embodiment of the present invention three;
Fig. 4 is the structural drawing of reconfigurable processing unit in the image processing apparatus of the mammary gland peripheral tissues equilibrium of the embodiment of the present invention four.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present application, be clearly and completely described the technical scheme in the embodiment of the present application, obviously, described embodiment is only some embodiments of the present application, instead of whole embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not making the every other embodiment obtained under creative work prerequisite, all belong to the scope of the application's protection.
Embodiment one
Be the process flow diagram of the image processing method of a kind of mammary gland peripheral tissues equilibrium of the embodiment of the present invention one with reference to figure 1, Fig. 1, the present embodiment method specifically can comprise:
Step 101, extracts human body and effectively organizes mask image from mammary gland original image.
In the present embodiment, first obtaining mammary gland partial original image, secondly, realizing effectively organizing mask image to extracting human body from this original image by the mode split original image.Be applicable to mammary gland partial original image by multiple method at present split, such as based on histogrammic O TSU dividing method, the dividing method such as method that increases based on region, because mammary gland partial original image tissue is distincter with background contrast, therefore, preferably, can adopt when the present embodiment specific implementation based on histogrammic O TSU dividing method.Concrete, the method for mammary gland partial original image segmentation is prior art, adopts which kind of method to realize extracting the enforcement that mask image does not all affect the present embodiment.
Step 102, carries out greyscale transformation to described original image and obtains pending image, decomposes described pending image and obtains a low-frequency image and multiple high frequency imaging.
Mammary gland original image refers to the view data without any process that image picking-up system collects, in image processing method conventional at present, all need to carry out the data that greyscale transformation obtains linearly changing with actual density, so that realize follow-up all data computings to original image.Here greyscale transformation can adopt log-transformation or other linear transformation process.
When specific implementation, multi-resolution pyramid algorithm can be adopted to decompose described pending image and to obtain a low-frequency image and multiple high frequency imaging.Multi-resolution pyramid algorithm applies a class algorithm more widely in Image Fusion, its principle is first by every width image pyramid representation of spatial registration, and on each equivalent layer, carry out the pyramid of all images to merge with certain fusion rule the pyramid obtaining synthesizing, then the inverse process reconstructed image that the pyramid pyramid of synthesis generates is obtained the image after processing.In the present embodiment, preferably laplacian pyramid algorithm can be adopted.
Concrete, this step is that the pending picture breakdown obtained through greyscale transformation is become a low-frequency image G n+1with multiple high frequency imaging S 1, S 2... S n.Wherein, N refers to the maximum number of plies that described multi-resolution pyramid algorithm decomposes pending image and adopts.From the 1st layer to the corresponding high frequency imaging of the every one deck of n-th layer, then after decomposing, obtain N number of high frequency imaging and a low-frequency image.
High frequency imaging S 1, S 2... S n, along with level increase picture size with 2 multiple successively successively decrease.The general enhancing of high-frequency information realization to different levels information by changing in different levels, therefore, be widely used at present based on multi-resolution pyramid algorithm in x line image process field, its level discharge rating process also becomes prior art.In the present embodiment, specifically can adopt the low-pass filter being of a size of 5*5, decomposing the maximum number of plies N adopted is 10.Certainly, can adopt other sizes and other maximum numbers of plies according to different demands, these do not affect the enforcement of the present embodiment yet.
Step 103, calculates grey scale mapping curve according to described mask image, described pending image and body of gland percentage composition.
Step 104, adopts grey scale mapping curve to carry out grey scale mapping to low-frequency image and obtains mapping rear low-frequency image.
Unify linear compression due to individual difference and can not adapt to all situations, if cross conference for compactness mammary gland peripheral tissues gray compression degree to change original body of gland distribution characteristics, and do not reach expected effect for sparse type mammary gland peripheral tissues gray compression degree is too small.Therefore the present invention dynamically generates grey scale mapping curve according to body of gland composition, utilizes this curve to carry out equilibrium to low-frequency image, to realize keeping primitive character for body of gland area grayscale as far as possible, adopts nonlinear mode progressively to increase suppressed range for peripheral tissues.
The present embodiment specifically matching can generate grey scale mapping curve by the following method, comprising:
The first step, according to the effective tissue regions histogram of described mask image statistics human body, according to described histogram, described body of gland percentage composition and preset peripheral tissues number percent definite threshold Tg and Ta, and according to body of gland percentage composition determination compressibility coefficient Kg and Ka; Wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
Second step, according to Kg, Ka, Tg, Ta and described pending gradation of image maximal value maxP, determines at least three unique points, and generates grey scale mapping curve according to described at least three unique point matchings.
When specific implementation, can determine that four unique points are as follows respectively:
P1 (0,0), P2 (Tg, Tg), P3 (Ta, (Ta-Tg) * Kg), P4 (maxP, (maxP-Ta) * Ka), from these four points, select three or four, adopt fitting of a polynomial algorithm or linear fit algorithm matching to generate grey scale mapping curve.
In practical operation, the specific implementation process of the first step is: statistics human body effective tissue regions grey level histogram Hist, according to body of gland percentage composition α and preset peripheral tissues number percent β (β refers to that skin tissue area number of pixels accounts for the number percent of whole ROI region number of pixels), alpha+beta should be not more than 1; Calculate gray threshold Tg and Ta according to body of gland percentage composition α and threshold value beta, tissue region being divided into three parts, is body of gland, fat, skin respectively, and wherein, Tg is used to the gray threshold splitting gland tissue and adipose tissue; Ta is used to the gray threshold splitting adipose tissue and skin histology.Also need to determine compressibility coefficient Kg and Ka according to body of gland percentage composition α, these two compressibility coefficient can be determined according to following table, and when number percent α is larger, compressibility coefficient Kg and Ka value should be larger.
Concrete, compressibility coefficient Kg and Ka can be determined by the corresponding relation of the body of gland percentage composition that pre-sets and compressibility coefficient Kg and Ka;
Concrete, Kg and Ka can be determined by the corresponding relation of the body of gland percentage composition shown in following form and compressibility coefficient.That is, the numerical range fallen into according to body of gland percentage composition α determines corresponding compressibility coefficient.
The concrete computation process of above-mentioned gray threshold Tg and Ta is as follows: first add up tissue area pixel number, and tissue area pixel sum M is multiplied by body of gland percentage composition α and obtains summer tissue and comprise number of pixels S g; S is multiplied by (1-β) obtains body of gland and adipose tissue regions comprises number of pixels S a; Setting an accumulator initial value is zero, then i starts from scratch traversal tissue area grayscale histogram Hist, this passage numerical value Hist [i] is added in totalizer sum total, when totalizer total starts to be greater than S through a grey level histogram passage totalizer gtime, record current i value for Tg; When total starts to be greater than S gtime, record current i value for Ta.
In practical operation, the specific implementation process of second step is: adopt fitting of a polynomial algorithm to carry out process of fitting treatment to four unique points and obtain a curve, using this curve as grey scale mapping curve.
Step 105, obtains high frequency imaging after gain to each high frequency imaging as gain process.
Due to gland tissue and peripheral tissues's contrast and noise variance all very large, therefore the intensity profile scope that grey scale mapping process can change low-frequency image is carried out to low-frequency image, but the improvement of its contrast to body of gland region is also not obvious, and peripheral tissues still possesses very strong contrast, therefore also need to carry out gain process to the high frequency imaging in different levels, to improve the phenomenon of image comparison skewness weighing apparatus.
When specific implementation, in advance one group of gain coefficient fixed can be set for all high frequency imagings, utilize these fixing gain coefficients to process the high-frequency information in different levels.The gain coefficient that high frequency imaging is generally set be greater than 1 numerical value.Certainly in order to improve high frequency imaging gain process effect better, can be the gain coefficient that every layer of high frequency imaging arranges dynamic change with reference to more factor.
Step 106, obtains the image after equilibrium to high frequency imaging after low-frequency image after described mapping and described gain as reconstruction processing.
It should be noted that, in the present embodiment, image reconstruction process is the same with normal Pyramid Reconstruction process, namely by being added with high frequency imaging again convolution after low-frequency image interpolation, successively process until be reconfigured to last one deck obtain equilibrium after image.
It should be noted that, sequencing requirement not strict between each step of the S101-S106 described in the present embodiment, when specific implementation, can according to the execution sequence shown in Fig. 1, also such as first S101 and S102 can be performed side by side according to other execution sequences, perform S103, S104 and S105 side by side again, finally perform S106, the present invention can certainly be realized according to other execution sequences.Can be found out by above-described embodiment, the body of gland composition matching that the present invention is based on image obtains grey scale mapping curve, utilizes this grey scale mapping curve to carry out mapping process to low-frequency image, like this, can reduce the gray difference of central area and peripheral tissues; Then again gain process is carried out to each layer of high frequency imaging, like this, make body of gland region and peripheral tissues's high-frequency information more balanced, finally, the image after obtaining equilibrium is reconstructed to the high frequency imaging after low-frequency image and gain process.Because grey scale mapping curve is based on the curve that obtains of body of gland composition matching, on the one hand, can under the prerequisite ensureing the distribution of body of gland primitive character, can the intensity of conservative control equilibrium; On the other hand, owing to considering the difference of image individuality, improve the consistance of result to a certain extent.In addition, because the present invention carries out multi-resolution decomposition to pending image, equilibrium treatment can be carried out under multiple yardstick, higher treatment effeciency can be ensured.
For the specific implementation of step 104 in above-described embodiment one, present invention also offers preferred version, be the gain coefficient that high frequency imaging arranges dynamic change, utilize the gain coefficient arranged to carry out gain process to high frequency imaging, to improve gain effect.This method for optimizing, comprising: step 1041 ~ 1043,
Step 1041, the low-frequency image half-tone information corresponding according to each high frequency imaging determines the first gain coefficient that each high frequency imaging is corresponding.
Concrete, determine the first gain coefficient in the following manner, comprising:
The low-frequency image corresponding according to every layer of high frequency imaging determines corresponding Tg and Ta, and wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
When specific implementation, also can directly adopt in above-described embodiment 1 about the threshold value Tg determined in grey scale mapping curve fitting process and Ta; But because in restructuring procedure, these two values may change, therefore all according to the low-frequency image of correspondence, corresponding these two threshold values of Tg and Ta are determined for each layer of high frequency imaging, its better effects if.
According to formula F gi=(FgN-Fa) * i/N+Fa, calculate the corresponding parameter Fgi in i-th layer of high frequency imaging body of gland region, wherein FgN is the corresponding parameter in n-th layer high frequency imaging body of gland region, and FgN span is 1 ~ 5, Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1;
According to the i-th floor height image characteristic point A1 (0 frequently, Fgi), A2 (Tg, Fgi), A3 (Ta, and A4 (maxP Fa), Fa) matching obtains the first gain trace, determines according to described first gain trace the first gain coefficient that i-th layer of high frequency imaging is corresponding.
Step 1042, determines according to the half-tone information of each high frequency imaging the second gain coefficient that each high frequency imaging is corresponding.
Concrete, determine the second gain coefficient in the following manner, comprising:
Calculate human body and effectively organize interior high-frequency gain maximal value max (abs (Si (ROI))), mean value ts1 and two times mean value ts2, and determine attenuation coefficient Fs, this attenuation coefficient Fs span is 0.5 ~ 1, and Fs is greater than Fa, Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1; Wherein, Si (ROI) is that the human body of the i-th floor high frequency imaging effectively organizes district;
According to the i-th floor height image characteristic point B1 (0,1), B2 (ts1,1), B3 (ts2, Fs) and B4 (max (abs (S frequently i(ROI))), Fa) matching obtains the second gain trace, determines according to described second gain trace the second gain coefficient that i-th layer of high frequency imaging is corresponding.
Step 1043, according to formula S i'=Gain2i* (Gain1i*S i-S i)+S i, calculate high frequency imaging after gain; Wherein, Gain2i represents that the second gain coefficient that i-th layer of high frequency imaging is corresponding, Gain1i represent the first gain coefficient that i-th layer of high frequency imaging is corresponding, S irepresent i-th layer of high frequency imaging, S i' represent i-th layer of high frequency imaging after gain, the span of i is 1,2 ... N-1, N, N refers to the maximum number of plies that described multi-resolution pyramid algorithm decomposes pending image and adopts.
The gain of the high frequency imaging determined due to said method considers picture contrast, noise profile feature, the increase of noise can't be caused while image body of gland region contrast is improved, use because disposal route can merge with multi-resolution decomposition Enhancement Method in addition, therefore efficiency is higher.
On the basis of above-described embodiment one, present invention also offers preferred scheme, the program is with the difference of the scheme of enforcement, adds artifact process, to reduce artifact effects in image further, to improve portfolio effect further in the process of reconstruction processing.Below by embodiment two, this preferred version is explained.
Embodiment two
Be the process flow diagram of the image processing method of a kind of mammary gland peripheral tissues equilibrium of the embodiment of the present invention two with reference to figure 2, Fig. 2, the present embodiment method specifically can comprise:
Step 201, extracts human body and effectively organizes mask image from mammary gland original image;
Step 202, carries out greyscale transformation to described original image and obtains pending image, decomposes described pending image and obtains a low-frequency image and multiple high frequency imaging;
Step 203, calculates grey scale mapping curve according to described mask image, described pending image and body of gland percentage composition;
Step 204, adopts grey scale mapping curve to carry out grey scale mapping to low-frequency image and obtains mapping rear low-frequency image;
Step 205, obtains high frequency imaging after gain to each high frequency imaging as gain process;
Step 201 ~ 205 are identical with step S101 ~ S105 in embodiment one, can describe accordingly, do not repeat them here in reference example one.Step 206, makes artifact to high frequency imaging after low-frequency image after mapping and gain and reconstruction processing obtains the image after equilibrium.
It should be noted that, sequencing requirement not strict between each step of the S201-S206 described in the present embodiment, when specific implementation, can according to the execution sequence of S201-S206, also such as first S201 and S202 can be performed side by side according to other execution sequences, perform S203, S204 and S205 side by side again, finally perform S206, the present invention can certainly be realized according to other execution sequences.
Concrete, this step S206 can realize in such a way, comprising:
Realize the reconstruction processing of i-th layer in the following manner successively, wherein i successively value be N, N-1, N-2 ... 2,1:
The low-frequency image corresponding to i-th layer of high frequency imaging is made interpolation convolution and is obtained low-frequency image GR i, then calculate i-th layer of high frequency imaging S after described low-frequency image GRi gain i' between with value GS i; Every i layer reconstruction result is as low-frequency image corresponding to the i-th-1 layer high frequency imaging;
Carry out down-sampledly obtaining down-sampled mask image to mask image, determine region, image border Edge according to down-sampled mask image i;
According to GS i* Edge i> I bithe outline region of marking image fringe region, wherein I biit is signature grey scale value in the background of the low-frequency image that i-th layer of high frequency imaging is corresponding; In background, signature grey scale value refers to mean value or the intermediate value of all gray scales of non-human tissue regions in image;
According to formula GPi=(GRi+ (k*Si ' * Edgei+Si ' * (1-Edgei))), wherein k span is 0 ~ 1, calculate i-th layer of reconstruction result GPi, again by reconstruction result GPi China and foreign countries contour area pixel assignment be Ib, realize the reconstruction processing of i-th layer successively until complete the 1st layer of reconstruction processing obtain equilibrium after image.
The method of the invention described above embodiment two be reconstruction processing in embodiment one basis on increase artifact process, inventor's analysis chart is the strongest as adjacent edges contrast, the artifact caused after image high frequency enhancement is the most obvious at adjacent edges, other regions are then very weak, therefore artifact process of the present invention is mainly near image border, to reduce artifact effects in image, to improve portfolio effect further.
Embodiment three
For realizing said method, present invention also offers the image processing apparatus of mammary gland peripheral tissues equilibrium.
Be the structural drawing of the image processing apparatus of a kind of mammary gland peripheral tissues equilibrium of the embodiment of the present invention three with reference to figure 3, Fig. 3, the present embodiment device specifically can comprise:
Extraction unit 301, effectively organizes mask image for extracting human body from mammary gland original image;
Resolving cell 302, obtains pending image for carrying out greyscale transformation to described original image, decomposes described pending image and obtains a low-frequency image and multiple high frequency imaging;
Grey scale mapping curve unit 303, for calculating grey scale mapping curve according to described mask image, described pending image and body of gland percentage composition;
Map unit 304, carries out grey scale mapping for adopting grey scale mapping curve to low-frequency image and obtains mapping rear low-frequency image;
Gain processing unit 305, for obtaining high frequency imaging after gain to each high frequency imaging as gain process;
Reconfigurable processing unit 306, for obtaining the image after equilibrium to high frequency imaging after low-frequency image after mapping and gain as reconstruction processing.
Preferably, described resolving cell obtains pending image specifically for carrying out greyscale transformation to described original image, adopts multi-resolution pyramid algorithm to decompose described pending image and obtains a low-frequency image and multiple high frequency imaging.Here greyscale transformation can adopt log-transformation or other linear transformations.
Preferably, described grey scale mapping curve unit, comprising:
Threshold value determination subelement, for adding up the effective tissue regions histogram of human body according to mask image, according to described histogram, described body of gland percentage composition and preset peripheral tissues number percent definite threshold Tg and Ta, and according to body of gland percentage composition determination compressibility coefficient Kg and Ka; Wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
Matching subelement, for according to Kg, Ka, Tg, Ta and described pending gradation of image maximal value maxP, determines at least three unique points, and generates grey scale mapping curve according to described at least three unique point matchings.
Concrete, described matching subelement, may be used for, according to Kg, Ka, Tg, Ta and described pending gradation of image maximal value maxP, determining four unique points, and generate grey scale mapping curve according to described four unique point matchings, wherein, described four unique points are respectively: P1 (0,0), P2 (Tg, Tg), P3 (Ta, (Ta-Tg) * Kg), P4 (maxP, (maxP-Ta) * Ka).
Described matching subelement can adopt fitting of a polynomial algorithm or linear fit algorithm to generate grey scale mapping curve.
Preferably, described gain processing unit, comprising:
First gain coefficient determination subelement, for determining according to the half-tone information of low-frequency image corresponding to each high frequency imaging the first gain coefficient that each high frequency imaging is corresponding;
Second gain coefficient determination subelement, for determining the second gain coefficient that each high frequency imaging is corresponding according to the half-tone information of each high frequency imaging;
Gain process subelement, for according to formula S i'=Gain2i* (Gain1i*S i-S i)+S i, calculate high frequency imaging after gain; Wherein, Gain2i represents that the second gain coefficient that i-th layer of high frequency imaging is corresponding, Gain1i represent the first gain coefficient that i-th layer of high frequency imaging is corresponding, S irepresent i-th layer of high frequency imaging, S i' represent i-th layer of high frequency imaging after gain, the span of i is 1,2 ... N-1, N, N refers to the maximum number of plies that described multi-resolution pyramid algorithm decomposes pending image and adopts.
Preferably, described first gain coefficient determining unit, comprising:
Intensity slicing value determination subelement, determine corresponding Tg and Ta for the low-frequency image corresponding according to every layer of high frequency imaging, wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
The corresponding parameter determination in region subelement, for according to formula F gi=(FgN-Fa) * i/N+Fa, calculate the corresponding parameter Fgi in i-th layer of high frequency imaging body of gland region, wherein FgN is the corresponding parameter in n-th layer high frequency imaging body of gland region, FgN span is 1 ~ 5, Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1;
First gain coefficient computation subunit, for according to the i-th floor height frequently image characteristic point A1 (0, Fgi), A2 (Tg, Fgi), A3 (Ta, and A4 (maxP Fa), Fa) matching obtains the first gain trace, according to described first gain trace as the first gain coefficient corresponding to i-th layer of high frequency imaging.
Preferably, described second gain coefficient determination subelement, comprising:
Parameter computation unit, interior high-frequency gain maximal value max (abs (Si (ROI))), mean value ts1 and two times mean value ts2 is effectively organized for calculating human body, and determine attenuation coefficient Fs, this attenuation coefficient Fs span is 0 ~ 1, and Fs is greater than Fa, Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1; Wherein, Si (ROI) is that the human body of the i-th floor high frequency imaging effectively organizes district;
Second gain coefficient computation subunit, for according to the i-th floor height frequently image characteristic point B1 (0,1), B2 (ts1,1), B3 (ts2, Fs) and B4 (max (abs (S i(ROI))), Fa) matching obtains the second gain trace, determines according to described second gain trace the second gain coefficient that i-th layer of high frequency imaging is corresponding.
Preferably, present invention also offers the image processing apparatus of another kind of mammary gland peripheral tissues equilibrium, from the difference of embodiment three, embodiment four is that reconfigurable processing unit is different, the identical description that all can refer to each unit in above-described embodiment three of other unit.
As shown in Figure 4, described reconfigurable processing unit, comprising:
With value computation subunit, make interpolation convolution for the low-frequency image corresponding to i-th layer of high frequency imaging and obtain low-frequency image GR i,calculate i-th layer of high frequency imaging S after described low-frequency image GRi and gain again i' between with value GS i, i-th layer of reconstruction result is as low-frequency image corresponding to the i-th-1 layer high frequency imaging; Wherein i successively value be N, N-1, N-2 ... 2,1;
Subelement 3061 is determined in region, image border, down-sampledly obtaining down-sampled mask image, determining region, image border Edge according to down-sampled mask image for carrying out mask image i;
Region, image border outline determination subelement 3062, for according to GS i* Edge i> I bithe outline region of marking image fringe region, wherein I biit is signature grey scale value in the background of the low-frequency image that i-th layer of high frequency imaging is corresponding;
Reconstruction processing subelement 3063, for according to formula
GPi=(GRi+ (k*Si ' * Edgei+Si ' * (1-Edgei))), wherein k span is 0 ~ 1, calculate i-th layer of reconstruction result GPi, again by reconstruction result GPi China and foreign countries contour area pixel assignment be Ibi, realize the reconstruction processing of i-th layer successively until complete the 1st layer of reconstruction processing obtain equilibrium after image.
Can be found out by above-described embodiment, the image processing apparatus of mammary gland peripheral tissues provided by the invention equilibrium, body of gland composition matching based on image obtains grey scale mapping curve, this grey scale mapping curve is utilized to carry out mapping process to low-frequency image, like this, the gray difference of central area and peripheral tissues can be reduced; Then again gain process is carried out to each layer of high frequency imaging, like this, make body of gland region and peripheral tissues's high-frequency information more balanced, finally, the image after obtaining equilibrium is reconstructed to the high frequency imaging after low-frequency image and gain process.Because grey scale mapping curve is based on the curve that obtains of body of gland composition matching, on the one hand, can under the prerequisite ensureing the distribution of body of gland primitive character, can the intensity of conservative control equilibrium; On the other hand, owing to considering the difference of image individuality, improve the consistance of result to a certain extent.In addition, because the present invention carries out multiple dimensioned decomposition to pending image, equilibrium treatment can be carried out under multiple yardstick, higher treatment effeciency can be ensured.
For device embodiment, because it corresponds essentially to embodiment of the method, so relevant part illustrates see the part of embodiment of the method.Device embodiment described above is only schematic, the wherein said unit illustrated as separating component or can may not be and physically separates, parts as unit display can be or may not be physical location, namely can be positioned at a place, or also can be distributed in multiple network element.Some or all of module wherein can be selected according to the actual needs to realize the object of the present embodiment scheme.Those of ordinary skill in the art, when not paying creative work, are namely appreciated that and implement.
It should be noted that, in this article, the such as relational terms of first and second grades and so on is only used for an entity or operation to separate with another entity or operational zone, and not necessarily requires or imply the relation that there is any this reality between these entities or operation or sequentially.And, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thus make to comprise the process of a series of key element, method, article or equipment and not only comprise those key elements, but also comprise other key elements clearly do not listed, or also comprise by the intrinsic key element of this process, method, article or equipment.When not more restrictions, the key element limited by statement " comprising ... ", and be not precluded within process, method, article or the equipment comprising described key element and also there is other identical element.
Above the image processing method of a kind of mammary gland peripheral tissues equilibrium that the embodiment of the present invention provides and device are described in detail, apply specific case herein to set forth principle of the present invention and embodiment, the explanation of above embodiment just understands method of the present invention and core concept thereof for helping; Meanwhile, for one of ordinary skill in the art, according to thought of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.

Claims (11)

1. an image processing method for mammary gland peripheral tissues equilibrium, is characterized in that, described method comprises:
From mammary gland original image, extract human body effectively organize mask image;
Greyscale transformation is carried out to described original image and obtains pending image, decompose described pending image and obtain a low-frequency image and multiple high frequency imaging;
Grey scale mapping curve is calculated according to described mask image, described pending image and body of gland percentage composition;
Adopt grey scale mapping curve to carry out grey scale mapping to low-frequency image to obtain mapping rear low-frequency image;
As gain process, high frequency imaging after gain is obtained to each high frequency imaging;
As reconstruction processing, the image after equilibrium is obtained to high frequency imaging after low-frequency image after described mapping and described gain.
2. method according to claim 1, is characterized in that, describedly calculates grey scale mapping curve according to mask image, pending image and body of gland percentage composition, comprising:
According to the effective tissue regions histogram of mask image statistics human body, according to described histogram, described body of gland percentage composition and preset peripheral tissues number percent definite threshold Tg and Ta, and according to body of gland percentage composition determination compressibility coefficient Kg and Ka; Wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
Determine at least three unique points according to Kg, Ka, Tg, Ta and described pending gradation of image maximal value maxP, and generate grey scale mapping curve according to described at least three unique point matchings.
3. method according to claim 1, is characterized in that, describedly obtains high frequency imaging after gain to each high frequency imaging as gain process, comprising:
The low-frequency image half-tone information corresponding according to each high frequency imaging determines the first gain coefficient that each high frequency imaging is corresponding;
The second gain coefficient that each high frequency imaging is corresponding is determined according to the half-tone information of each high frequency imaging;
According to formula S i'=Gain2i* (Gain1i*S i-S i)+S i, calculate high frequency imaging after gain; Wherein, Gain2i represents that the second gain coefficient that i-th layer of high frequency imaging is corresponding, Gain1i represent the first gain coefficient that i-th layer of high frequency imaging is corresponding, S irepresent i-th layer of high frequency imaging, S i' represent i-th layer of high frequency imaging after gain, the span of i is 1,2 ... N-1, N, N refers to the maximum number of plies that described multi-resolution pyramid algorithm adopts when decomposing pending image.
4. method according to claim 3, is characterized in that, the half-tone information of the described low-frequency image corresponding according to each high frequency imaging is determined to comprise the first gain coefficient that each high frequency imaging is corresponding:
The low-frequency image corresponding according to every layer of high frequency imaging determines corresponding Tg and Ta, and wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
According to formula F gi=(FgN-Fa) * i/N+Fa, calculate the corresponding parameter Fgi in i-th layer of high frequency imaging body of gland region, wherein FgN is the corresponding parameter in n-th layer high frequency imaging body of gland region, and FgN span is 1 ~ 5, Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1;
According to the i-th floor height image characteristic point A1 (0 frequently, Fgi), A2 (Tg, Fgi), A3 (Ta, and A4 (maxP Fa), Fa) matching obtains the first gain trace, determines according to described first gain trace the first gain coefficient that i-th layer of high frequency imaging is corresponding.
5. method according to claim 3, is characterized in that, the described half-tone information according to each high frequency imaging is determined to comprise the second gain coefficient that each high frequency imaging is corresponding:
Calculate human body and effectively organize interior high-frequency gain maximal value max (abs (Si (ROI))), mean value ts1 and two times mean value ts2, and determine attenuation coefficient Fs, this attenuation coefficient Fs span is 0.5 ~ 1, and Fs is greater than Fa; Fa is the corresponding parameter in peripheral tissues region, and Fa span is 0.5 ~ 1, and wherein, Si (ROI) is that the human body of the i-th floor high frequency imaging effectively organizes district;
According to the i-th floor height image characteristic point B1 (0,1), B2 (ts1,1), B3 (ts2, Fs) and B4 (max (abs (S frequently i(ROI))), Fa) calculate the second gain trace, determine according to described second gain trace the second gain coefficient that i-th layer of high frequency imaging is corresponding.
6. method according to claim 1, is characterized in that, describedly obtains the image after equilibrium to high frequency imaging after low-frequency image after described mapping and described gain as reconstruction processing, comprising:
Realize the reconstruction processing of i-th layer in the following manner successively, wherein i successively value be N, N-1, N-2 ... 2,1:
The low-frequency image corresponding to i-th layer of high frequency imaging is made interpolation convolution and is obtained low-frequency image GR i, then calculate i-th layer of high frequency imaging S after described low-frequency image GRi gain i' between with value GS i; Every i layer reconstruction result is as low-frequency image corresponding to the i-th-1 layer high frequency imaging;
Carry out down-sampledly obtaining down-sampled mask image to mask image, determine region, image border Edge according to down-sampled mask image i;
According to GS i* Edge i> I bithe outline region of marking image fringe region, wherein I biit is signature grey scale value in the background of the low-frequency image that i-th layer of high frequency imaging is corresponding;
According to formula GP i=(GR i+ (k*S i' * Edge i+ S i' * (1-Edge i))), wherein k span is 0 ~ 1, calculates i-th layer of reconstruction result GP i, then by reconstruction result GP ithe pixel assignment of China and foreign countries' contour area is I bi, realize the reconstruction processing of i-th layer successively until complete the 1st layer of reconstruction processing obtain equilibrium after image.
7. method according to claim 1, is characterized in that, the described pending image of described decomposition obtains a low-frequency image and multiple high frequency imaging is specially:
Adopt multi-resolution pyramid algorithm to decompose described pending image and obtain a low-frequency image and multiple high frequency imaging.
8. an image processing apparatus for mammary gland peripheral tissues equilibrium, is characterized in that, described device comprises:
Extraction unit, effectively organizes mask image for extracting human body from mammary gland original image;
Resolving cell, obtains pending image for carrying out greyscale transformation to described original image, decomposes described pending image and obtains a low-frequency image and multiple high frequency imaging;
Grey scale mapping curve unit, for calculating grey scale mapping curve according to described mask image, described pending image and body of gland percentage composition;
Map unit, carries out grey scale mapping for adopting grey scale mapping curve to low-frequency image and obtains mapping rear low-frequency image;
Gain processing unit, for obtaining high frequency imaging after gain to each high frequency imaging as gain process;
Reconfigurable processing unit, for obtaining the image after equilibrium to high frequency imaging after low-frequency image after described mapping and described gain as reconstruction processing.
9. device according to claim 8, is characterized in that, described grey scale mapping curve unit, comprising:
Threshold value determination subelement, for adding up the effective tissue regions histogram of human body according to mask image, according to described histogram, described body of gland percentage composition and preset peripheral tissues number percent definite threshold Tg and Ta, and according to body of gland percentage composition determination compressibility coefficient Kg and Ka; Wherein, Tg represents gland tissue and adipose tissue intensity slicing value, and Ta represents adipose tissue and skin histology intensity slicing value;
Matching subelement, for determining at least three unique points according to Kg, Ka, Tg, Ta and described pending gradation of image maximal value maxP, and generates grey scale mapping curve according to described at least three unique point matchings.
10. device according to claim 8, is characterized in that, described gain processing unit, comprising:
First gain coefficient determination subelement, determines for the low-frequency image half-tone information corresponding according to each high frequency imaging the first gain coefficient that each high frequency imaging is corresponding;
Second gain coefficient determination subelement, for determining the second gain coefficient that each high frequency imaging is corresponding according to the half-tone information of each high frequency imaging;
Gain process subelement, for according to formula S i'=Gain2i* (Gain1i*S i-S i)+S i, calculate high frequency imaging after gain; Wherein, Gain2i represents that the second gain coefficient that i-th layer of high frequency imaging is corresponding, Gain1i represent the first gain coefficient that i-th layer of high frequency imaging is corresponding, S irepresent i-th layer of high frequency imaging, S i' represent i-th layer of high frequency imaging after gain, the span of i is 1,2 ... N-1, N, N refers to the maximum number of plies that described multi-resolution pyramid algorithm adopts when decomposing pending image.
11. devices according to claim 8, is characterized in that, described reconfigurable processing unit, comprising:
With value computation subunit, make interpolation convolution for the low-frequency image corresponding to i-th layer of high frequency imaging and obtain low-frequency image GR i, then calculate i-th layer of high frequency imaging S after described low-frequency image GRi and gain i' between with value GS i, every i layer reconstruction result is as low-frequency image corresponding to the i-th-1 layer high frequency imaging; Wherein i successively value be N, N-1, N-2 ... 2,1;
Subelement is determined in region, image border, down-sampledly obtaining down-sampled mask image, determining region, image border Edge according to down-sampled mask image for carrying out mask image i;
Region, image border outline determination subelement, for according to GS i* Edge i> I bithe outline region of marking image fringe region, wherein I biit is signature grey scale value in the background of the low-frequency image that i-th layer of high frequency imaging is corresponding;
Reconstruction processing subelement, for according to formula GP i=(GR i+ (k*S i' * Edge i+ S i' * (1-Edge i))), wherein k span is 0 ~ 1, calculates i-th layer of reconstruction result GP i, then by reconstruction result GP ithe pixel assignment of China and foreign countries' contour area is I bi, realize the reconstruction processing of i-th layer successively until complete the 1st layer of reconstruction processing obtain equilibrium after image.
CN201410707789.4A 2014-11-27 2014-11-27 A kind of mammary gland peripheral tissues balanced image processing method and device Active CN104574361B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410707789.4A CN104574361B (en) 2014-11-27 2014-11-27 A kind of mammary gland peripheral tissues balanced image processing method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410707789.4A CN104574361B (en) 2014-11-27 2014-11-27 A kind of mammary gland peripheral tissues balanced image processing method and device

Publications (2)

Publication Number Publication Date
CN104574361A true CN104574361A (en) 2015-04-29
CN104574361B CN104574361B (en) 2017-12-29

Family

ID=53090333

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410707789.4A Active CN104574361B (en) 2014-11-27 2014-11-27 A kind of mammary gland peripheral tissues balanced image processing method and device

Country Status (1)

Country Link
CN (1) CN104574361B (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105488765A (en) * 2015-11-23 2016-04-13 沈阳东软医疗系统有限公司 Image dynamic range adjustment method and apparatus
CN105701796A (en) * 2015-12-31 2016-06-22 上海联影医疗科技有限公司 Thickness balancing method and device for breast image and mammography system
WO2017059605A1 (en) * 2015-10-07 2017-04-13 南京巨鲨显示科技有限公司 Display grayscale curve correction system and method for mammary gland molybdenum target image
CN107292815A (en) * 2017-06-14 2017-10-24 上海联影医疗科技有限公司 Processing method, device and the breast imaging equipment of galactophore image
CN107316291A (en) * 2017-06-14 2017-11-03 上海联影医疗科技有限公司 Galactophore image processing method and breast imaging equipment
CN107862699A (en) * 2017-09-22 2018-03-30 中国科学院深圳先进技术研究院 Bone edges extracting method, device, equipment and the storage medium of Bone CT image
CN108009992A (en) * 2016-10-28 2018-05-08 北京东软医疗设备有限公司 A kind of method and apparatus of Image Adjusting
WO2018227943A1 (en) * 2017-06-14 2018-12-20 Shenzhen United Imaging Healthcare Co., Ltd. System and method for image processing
US10290108B2 (en) 2015-12-31 2019-05-14 Shanghai United Imaging Healthcare Co., Ltd. Methods and systems for image processing
CN110838091A (en) * 2019-10-25 2020-02-25 北京中电慧视科技有限公司 Fully self-adaptive infrared image contrast enhancement method and system
CN111311605A (en) * 2020-01-15 2020-06-19 深圳市安健科技股份有限公司 X-ray image segmentation method and terminal
US10949950B2 (en) 2017-06-14 2021-03-16 Shanghai United Imaging Healthcare Co., Ltd. System and method for image processing
CN113724144A (en) * 2020-05-22 2021-11-30 北京小米移动软件有限公司 Image processing method and image signal processor on terminal equipment
CN113892967A (en) * 2021-09-07 2022-01-07 青岛海信医疗设备股份有限公司 Ultrasonic image processing method and ultrasonic instrument

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5528703A (en) * 1992-02-18 1996-06-18 Neopath, Inc. Method for identifying objects using data processing techniques
CN102142133A (en) * 2011-04-19 2011-08-03 西安电子科技大学 Mammary X-ray image enhancement method based on non-subsampled Directionlet transform and compressive sensing
CN102456222A (en) * 2010-10-29 2012-05-16 深圳迈瑞生物医疗电子股份有限公司 Method and device for organized equalization in image
CN102890817A (en) * 2011-07-22 2013-01-23 深圳市蓝韵实业有限公司 Method and system for processing medical images

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5528703A (en) * 1992-02-18 1996-06-18 Neopath, Inc. Method for identifying objects using data processing techniques
CN102456222A (en) * 2010-10-29 2012-05-16 深圳迈瑞生物医疗电子股份有限公司 Method and device for organized equalization in image
CN102142133A (en) * 2011-04-19 2011-08-03 西安电子科技大学 Mammary X-ray image enhancement method based on non-subsampled Directionlet transform and compressive sensing
CN102890817A (en) * 2011-07-22 2013-01-23 深圳市蓝韵实业有限公司 Method and system for processing medical images

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
丰国栋: "数字化X线摄影图像增强方法研究", 《中国博士学位论文全文数据库 信息科技辑》 *

Cited By (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10467738B2 (en) 2015-10-07 2019-11-05 Nanjing Jusha Display Technology Co., Ltd. Display gray scale curve correction system and method for molybdenum target mammography
WO2017059605A1 (en) * 2015-10-07 2017-04-13 南京巨鲨显示科技有限公司 Display grayscale curve correction system and method for mammary gland molybdenum target image
CN105488765B (en) * 2015-11-23 2018-12-07 沈阳东软医疗系统有限公司 A kind of dynamic range of images method of adjustment and device
CN105488765A (en) * 2015-11-23 2016-04-13 沈阳东软医疗系统有限公司 Image dynamic range adjustment method and apparatus
CN105701796A (en) * 2015-12-31 2016-06-22 上海联影医疗科技有限公司 Thickness balancing method and device for breast image and mammography system
US11880978B2 (en) 2015-12-31 2024-01-23 Shanghai United Imaging Healthcare Co., Ltd. Methods and systems for image processing
US11049254B2 (en) 2015-12-31 2021-06-29 Shanghai United Imaging Healthcare Co., Ltd. Methods and systems for image processing
CN105701796B (en) * 2015-12-31 2018-09-18 上海联影医疗科技有限公司 The thickness equalization methods and device of breast image, mammography system
US10290108B2 (en) 2015-12-31 2019-05-14 Shanghai United Imaging Healthcare Co., Ltd. Methods and systems for image processing
CN108009992B (en) * 2016-10-28 2020-06-16 北京东软医疗设备有限公司 Image adjusting method and device
CN108009992A (en) * 2016-10-28 2018-05-08 北京东软医疗设备有限公司 A kind of method and apparatus of Image Adjusting
CN110832540A (en) * 2017-06-14 2020-02-21 上海联影医疗科技有限公司 Image processing system and method
US11562469B2 (en) 2017-06-14 2023-01-24 Shanghai United Imaging Healthcare Co., Ltd. System and method for image processing
CN107292815A (en) * 2017-06-14 2017-10-24 上海联影医疗科技有限公司 Processing method, device and the breast imaging equipment of galactophore image
US10949950B2 (en) 2017-06-14 2021-03-16 Shanghai United Imaging Healthcare Co., Ltd. System and method for image processing
CN107316291A (en) * 2017-06-14 2017-11-03 上海联影医疗科技有限公司 Galactophore image processing method and breast imaging equipment
CN110832540B (en) * 2017-06-14 2023-08-01 上海联影医疗科技股份有限公司 Image processing system and method
WO2018227943A1 (en) * 2017-06-14 2018-12-20 Shenzhen United Imaging Healthcare Co., Ltd. System and method for image processing
CN107862699A (en) * 2017-09-22 2018-03-30 中国科学院深圳先进技术研究院 Bone edges extracting method, device, equipment and the storage medium of Bone CT image
CN110838091A (en) * 2019-10-25 2020-02-25 北京中电慧视科技有限公司 Fully self-adaptive infrared image contrast enhancement method and system
CN111311605B (en) * 2020-01-15 2023-07-04 深圳市安健科技股份有限公司 X-ray image segmentation method and terminal
CN111311605A (en) * 2020-01-15 2020-06-19 深圳市安健科技股份有限公司 X-ray image segmentation method and terminal
CN113724144A (en) * 2020-05-22 2021-11-30 北京小米移动软件有限公司 Image processing method and image signal processor on terminal equipment
CN113892967A (en) * 2021-09-07 2022-01-07 青岛海信医疗设备股份有限公司 Ultrasonic image processing method and ultrasonic instrument
CN113892967B (en) * 2021-09-07 2023-10-27 青岛海信医疗设备股份有限公司 Ultrasonic image processing method and ultrasonic instrument

Also Published As

Publication number Publication date
CN104574361B (en) 2017-12-29

Similar Documents

Publication Publication Date Title
CN104574361A (en) Image processing method and device for mammary peripheral tissue equalization
CN107358586B (en) Image enhancement method, device and equipment
CN101727658B (en) Image processing method and device
EP3224800B1 (en) Simulating dose increase by noise model based multi scale noise reduction
CN110517327B (en) Underwater image enhancement method based on color correction and contrast stretching
Nguyen et al. Bounded self-weights estimation method for non-local means image denoising using minimax estimators
CN108876733B (en) Image enhancement method, device, equipment and storage medium
Kanwal et al. Region based adaptive contrast enhancement of medical X-ray images
CN110889812B (en) Underwater image enhancement method for multi-scale fusion of image characteristic information
TWI433053B (en) Method and system for image sharpness enhancement based on local feature of the image
CN110533614B (en) Underwater image enhancement method combining frequency domain and airspace
CN110322404B (en) Image enhancement method and system
Borges et al. Pipeline for effective denoising of digital mammography and digital breast tomosynthesis
Deng A generalized logarithmic image processing model based on the gigavision sensor model
CN104680485A (en) Method and device for denoising image based on multiple resolutions
CN104574284A (en) Digital X-ray image contrast enhancement processing method
EP1422662B1 (en) Image processing apparatus and method for contrast enhancement by histogram equalization
EP2232434A1 (en) Method of generating a multiscale contrast enhanced image.
CN111507912B (en) Mammary gland image enhancement method and device and computer readable storage medium
Agaian et al. Novel medical image enhancement algorithms
Sahoo Biomedical image fusion and segmentation using glcm
JP4072491B2 (en) Image processing apparatus, image processing method, program, and computer-readable medium
Malik et al. Contrast enhancement and smoothing of CT images for diagnosis
CN104978720A (en) Video image raindrop removal method and apparatus
Batna Comparative study between wavelet thresholding techniques (hard, soft and invariant-translation) in ultrasound images

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CP03 Change of name, title or address
CP03 Change of name, title or address

Address after: 110179 No. 177-1 Innovation Road, Hunnan District, Shenyang City, Liaoning Province

Patentee after: DongSoft Medical System Co., Ltd.

Address before: Hunnan New Century Road 110179 Shenyang city of Liaoning Province, No. 16

Patentee before: Dongruan Medical Systems Co., Ltd., Shenyang