CN106910179A - Multimode medical image fusion method based on wavelet transformation - Google Patents
Multimode medical image fusion method based on wavelet transformation Download PDFInfo
- Publication number
- CN106910179A CN106910179A CN201710046867.4A CN201710046867A CN106910179A CN 106910179 A CN106910179 A CN 106910179A CN 201710046867 A CN201710046867 A CN 201710046867A CN 106910179 A CN106910179 A CN 106910179A
- Authority
- CN
- China
- Prior art keywords
- frequency
- image
- component
- wavelet
- low
- 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
Links
- 230000009466 transformation Effects 0.000 title claims abstract description 30
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 17
- 238000000034 method Methods 0.000 claims abstract description 25
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 20
- 230000003044 adaptive effect Effects 0.000 claims abstract description 12
- 210000000056 organ Anatomy 0.000 claims description 30
- 230000004927 fusion Effects 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 7
- 230000015556 catabolic process Effects 0.000 claims description 6
- 201000007094 prostatitis Diseases 0.000 claims description 6
- 230000002068 genetic effect Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000005192 partition Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 210000001519 tissue Anatomy 0.000 description 3
- 210000003484 anatomy Anatomy 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 230000001575 pathological effect Effects 0.000 description 2
- 238000012797 qualification Methods 0.000 description 2
- 210000004872 soft tissue Anatomy 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 230000002308 calcification Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000002224 dissection Methods 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 230000001613 neoplastic effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000002285 radioactive effect Effects 0.000 description 1
- 238000005549 size reduction Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
- 210000001835 viscera Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/20064—Wavelet transform [DWT]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
A kind of multimode medical image fusion method based on adaptive wavelet packet transform, comprises the following steps：S1, wavelet transformation is carried out to the medical image of different modalities using adaptive wavelet packet transform wave filter, earlier figures picture is separately disassembled into the component that high frequency, low frequency and low-and high-frequency are combined；S2, the component of the high frequency, low frequency and low-and high-frequency combination obtained to the decomposition of any two width different modalities medical image are overlapped, and obtain the component that high frequency, low frequency and the low-and high-frequency of fused images are combined；The component that S3, the high frequency to fused images, low frequency and low-and high-frequency are combined carries out discrete wavelet inverse transformation, obtains the fused images of original size.Be applied to wavelet transformation technique in image processing field by the present invention, adaptive wavelet packet transform can effective details susceptibility of the Lifting Wavelet coefficient to low frequency component, so that it is more accurate further to decompose the high frequency that obtains, low frequency component, and then cause that syncretizing effect gets a promotion.
Description
Technical field
The present invention relates to medical field, in particular by the side that image processing techniques is merged to multi-modality medical image
Method, specifically a kind of multimode medical image fusion method based on wavelet transformation.
Background technology
At present, with the fast development of computer science and technology and medical treatment influence engineering science, many elder generations are occurred in that in the world
The medical imaging device for entering, the medical image of multiple modalities is provided for clinical medicine diagnosis, and these images are anti-from different aspect
The different information of organization of human body, internal organs and pathological tissues are reflected.
Such as CT (Computed Tomography) image has stronger spatial resolution and geometrical property, to bone
Imaging it is very clear, it can provide preferable reference to lesion localization, but then relatively low to the contrast of soft tissue.MR
(Magnetic Resonance) image can clearly reflect the anatomical structure of soft tissue, organ, blood vessel etc., be conducive to determining
Focus scope, but MR images are insensitive to calcification point, and geometric distortion can be occurred by magnetic disturbance.SPEC, PET image can be obtained
Radioactive concentration to the arbitrarily angled fault plane of human body is distributed, and can reflect the metaboilic level and blood flow state of histoorgan, right
Neoplastic lesion is presented " focus ", there is provided the function information of human body, but their resolution ratio is poor, hardly results in accurate dissection and ties
Structure, is not easy to differentiate tissue, the border of organ.As can be seen here, different imaging techniques have the advantage of itself also while possessing one
A little limitation, these images be to the form and function information obtained by the same anatomical structure of human body each other difference, complement one another
's.
In clinical diagnosis, the image of single mode tends not to the enough information required for providing doctor, therefore, if
The medical image of different modalities can be carried out appropriate fusion, anatomic information and function information is organically combined, one
The information from various imaging sources is synthetically expressed simultaneously on width image, the comprehensive feelings of pathological tissues or organ are understood so as to doctor
Condition, and make more accurate diagnosis or make the therapeutic scheme of more scientific optimization, this will promote modern medicine clinical
The huge advance of technology.
The characteristics of for different modalities medical image, in order to the important information of preferably fusion different images, we
Propose the multimode medical image fusion method based on adaptive wavelet packet transform.On the premise of image energy feature calculation, from
Adapting to selection optimal wavelet coefficient carries out wavelet transformation, so that wavelet transformation can effectively by the background and details of image
Make a distinction, under adjustable weight, the high and low frequency information of two secondary different modalities images is folded according to different proportion
Plus, fast and efficiently complete fusion process.
The content of the invention
The purpose of the present invention is directed to the problem of image co-registration, proposes a kind of multi-modality medical image based on wavelet transformation
Fusion method.
The technical scheme is that：
A kind of multimode medical image fusion method based on adaptive wavelet packet transform, it is comprised the following steps：
S1, wavelet transformation is carried out to the medical image of different modalities using adaptive wavelet packet transform wave filter, by earlier figures picture
It is separately disassembled into the component that high frequency, low frequency and low-and high-frequency are combined；
S2, the component of the high frequency, low frequency and low-and high-frequency combination obtained to the decomposition of any two width different modalities medical image enter
Row superposition, obtains the component that high frequency, low frequency and the low-and high-frequency of fused images are combined；
The component that S3, the high frequency to fused images, low frequency and low-and high-frequency are combined carries out discrete wavelet inverse transformation, obtains original
The fused images of size.
In step S1 of the invention, low frequency component is decomposed again according to yardstick demand.
Step S1 of the invention is specially：
S1.1, wavelet structure base filter bank, i.e., based on meeting orthogonality condition fixed length wavelet filter, different
Multiple filter coefficients are obtained under initial condition and builds wavelet basis filter bank, one is randomly choosed in wavelet basis filter bank
Filter coefficient is used as initialization wave filter；
S1.2, using initialization wave filter the medical image of multiple mode is carried out wavelet transformation obtain ground floor decompose point
Amount, i.e., obtain the low frequency in image low frequency component LL1 in the horizontal and vertical directions, horizontal direction by wavelet transform
With the high fdrequency component LH1 in vertical direction, the low frequency component HL1 on the high frequency and vertical direction in horizontal direction and level and
High fdrequency component HH1 in vertical direction；
The fluctuation of S1.3, the Energy distribution according to image and texture selects filter coefficient again in wavelet basis filter bank
To step S1.2 obtain ground floor decomposed component in low frequency component LL1 carry out second wavelet transformation, obtain LL2, HH2,
HL2, LH2 repeat step S1.3 several times, complete picture breakdown.
In step S1.3 of the invention, the number of times that repeat step S1.3 carries out picture breakdown is 3-5 times.
In step S1.3 of the invention, Energy distribution and texture according to image fluctuate and are selected in wavelet basis filter bank
The specific method of filter coefficient is：
S1.3-1, the step of carry out organ to image and extract：
Four components in current decomposition component are processed, human organ is extracted from the background of black,
It is specific that human organ is extracted using image partition method, or human organ image is extracted using the method for setting boundary threshold
HLo and LHo；
The step of S1.3-2, calculating image texture characteristic parameter：
Horizontal image least energy Eh, the image least energy Ev of longitudinal direction, horizontal image are calculated using following formula
Minimum extreme difference Ep, image minimum extreme difference Eq, the transverse direction lines maximum fluctuation value Em of longitudinal direction and
Longitudinal lines maximum fluctuation value En；
Eh=min (max (| HLo |)) (1)
Ev=min (max (| LHo |)) (2)
Ep=min (max (| HLo |)-min (| HLo |)) (3)
Eq=min (max (| LHo |)-min (| LHo |)) (4)
Wherein, HLo represents the human body device of the low frequency extraction corresponding with the high fdrequency component LH1 in vertical direction in horizontal direction
Official's image, LHo represent the human organ figure of the high frequency extraction corresponding with the low frequency component HL1 in vertical direction in horizontal direction
Picture；I, j represent the sequence number of row, and i ', j ' represent the sequence number of row；S1.3-3, according to foregoing six image texture characteristic parameter acquirings
Wavelet basis filter coefficient.
In step S1.3-3 of the invention, wavelet basis filter coefficient acquisition methods are：Obtained according to step S1.3-2
Six image texture characteristic parameters, wavelet basis filter coefficient, preferably genetic programming algorithm are obtained using optimization algorithm.
In step S1.3-3 of the invention, wavelet basis filter coefficient acquisition methods are：A certain amount of sample image is extracted,
The table of comparisons of six characteristic indexs and wavelet basis filter coefficient is set up, when calculating in real time, is obtained according to step S1.3-2
Six image texture characteristic parameter controls are searched and obtain corresponding wavelet basis filter coefficient.
In step S1.3-1 of the invention, human organ is extracted using the method for setting boundary threshold and specifically includes following step
Suddenly：
To when high fdrequency component LHn, the horizontal direction on the previous low frequency and vertical direction decomposed in the horizontal direction for obtaining
On high frequency and vertical direction on two subgraphs of low frequency component HLn carry out operations described below, respectively obtain organic image HLo
And LHo：
As sum (| Row |)>During Threshold1, retain current line pixel data, otherwise, current line pixel data is made
It is background, deletes the row data；
As sum (| Column |)>During Threshold2, retain and work as prostatitis pixel data, otherwise, prostatitis pixel data will be worked as
As background, the row data are deleted；
Wherein, Row is capable pixel, and Column is the pixel of row, and Threshold1, Threshold2 are respectively organ moulds
The corresponding parameter threshold of type.
In step S2 of the invention, two width different modalities medical images are decomposed and obtains the combination of high frequency, low frequency and low-and high-frequency
Component be overlapped under certain weight, obtain the component that the high frequency of fused images, low frequency and low-and high-frequency are combined, specially：
S2.1, the subgraph of any two mode medical image after decomposition is merged with certain weight, according to
The principle that low-and high-frequency correspondence is added is carried out, wherein the arbitrary small number between the fusion weight W selections 0.1~0.9 of a sub-picture, separately
Outer one secondary fusion weight is 1-W；
S2.2, the decomposition result AHHn, ALLn, AHLn, ALHn and BHHn to different modalities image A and B, BLLn, BHLn,
BLHn is overlapped computing, and n represents n-th layer decomposition, obtains four component Fs HHn, FLLn, FHLn and the FLHn of fused images F：
FHHn=wHH*AHHn+ (1-wHH) * BHHn；WHH is that high frequency merges weight
FLLn=wLL*ALLn+ (1-wLL) * BLLn；WLL is that low frequency merges weight
FHLn=wHL*AHLn+ (1-wHL) * BHLn；WHL is that low-and high-frequency merges weight
FLHn=wLH*ALHn+ (1-wLH) * BLHn；WLH is that low high frequency merges weight.
Beneficial effects of the present invention：
Be applied to wavelet transformation technique in image processing field by the present invention, and adaptive wavelet packet transform can effective Lifting Wavelet
Details susceptibility of the coefficient to low frequency component so that it is more accurate further to decompose the high frequency that obtains, low frequency component, and then causes
Syncretizing effect gets a promotion；The process of the adaptive wavelet packet transform coefficient analysis carried out for medical image feature, has distinguished organ
Image and invalid background, effectively improve the reliability of wavelet coefficient, and the method can form effectively combination with medical big data,
Robustness of the formation based on organ model stronger medical image segmentation, analysis model.
Brief description of the drawings
Fig. 1 is structural representation of the invention.
Fig. 2 is that image energy feature of the invention compares schematic diagram with wavelet coefficient.
Fig. 3 is that the organ of medical image and the schematic diagram of background are quickly distinguished using threshold decision in the present invention.
Fig. 4 is the image schematic diagram that texture analysis is obtained in the present invention.
Specific embodiment
The present invention is further illustrated with reference to the accompanying drawings and examples.
As Figure 1-4, a kind of multimode medical image fusion method based on adaptive wavelet packet transform, it includes following step
Suddenly：
S1, wavelet transformation is carried out to the medical image of different modalities using adaptive wavelet packet transform wave filter, by earlier figures picture
It is separately disassembled into the component that high frequency, low frequency and low-and high-frequency are combined；
S2, the component of the high frequency, low frequency and low-and high-frequency combination obtained to the decomposition of any two width different modalities medical image enter
Row superposition, obtains the component that high frequency, low frequency and the low-and high-frequency of fused images are combined；
The component that S3, the high frequency to fused images, low frequency and low-and high-frequency are combined carries out discrete wavelet inverse transformation, obtains original
The fused images of size；
In step S1 of the invention, low frequency component is decomposed again according to yardstick demand.
Step S1 of the invention is specially：
S1.1, wavelet structure base filter bank, i.e., based on meeting orthogonality condition fixed length wavelet filter, different
Multiple filter coefficients are obtained under initial condition and builds wavelet basis filter bank, one is randomly choosed in wavelet basis filter bank
Filter coefficient is used as initialization wave filter；
S1.2, using initialization wave filter the medical image of multiple mode is carried out wavelet transformation obtain ground floor decompose point
Amount, i.e., obtain image both horizontally and vertically by wavelet transform (Discrete Wavelet Transform, DWT)
On low frequency component LL1, the high fdrequency component LH1 on low frequency and vertical direction in horizontal direction, the high frequency in horizontal direction and
Low frequency component HL1 in the vertical direction and high fdrequency component HH1 for both horizontally and vertically going up；
The fluctuation of S1.3, the Energy distribution according to image and texture selects filter coefficient again in wavelet basis filter bank
To step S1.2 obtain ground floor decomposed component in low frequency component LL1 carry out second wavelet transformation, obtain LL2, HH2,
HL2, LH2 repeat step S1.3 several times, complete picture breakdown.
In step S1.3 of the invention, repeat step S1.3 carries out the number of times of picture breakdown for 3-5 time, syncretizing effect with
Equalization point can be reached in computational efficiency, due to each exploded view as size reduction is the 1/4 of current layer, so stock size is very
Big image can just use more than 5 layers of decomposition and real-time is not high, and less than 3 layers of syncretizing effect can be relatively weaker；For
Large-size image can select 5 layers of decomposition, 3 layers of decomposition of reduced size image selection, it is also possible to select 4 layers to decompose as
Equalization point.
According to medical image energy statisticses：Because medical image is usually gray level image, background is with noisy black
Background, prospect is the human organ for needing to be observed, and in step S1.3, Energy distribution and texture according to image fluctuate small
The specific method of selection filter coefficient is in ripple base filter bank：,
S1.3-1, the step of carry out organ to image and extract：
Four components in current decomposition component are processed, human organ is extracted from the background of black,
It is specific that human organ is extracted using image partition method, or human organ image is extracted using the method for setting boundary threshold
HLo and LHo；
The step of S1.3-2, calculating image texture characteristic parameter：
Horizontal image least energy Eh, the image least energy Ev of longitudinal direction, horizontal image are calculated using following formula
Minimum extreme difference Ep, image minimum extreme difference Eq, the transverse direction lines maximum fluctuation value Em and longitudinal lines maximum fluctuation value En of longitudinal direction；
(image energy parameter Eh, Ev are calculated, qualifications are set to the sub- energy minimum of the high pass of image, enable to small echo base system
Unity and coherence in writing feature instantiation out, is improved the sensitiveness to detailed information by number；The wavelet coefficient obtained after the static decomposition of image is maximum
Value Ep, Eq, to the wavelet coefficient distribution of downscaled images；Horizontal, longitudinal direction lines maximum fluctuation Em, En are image row, column ash
The relation of the difference of angle value sum, reflection wavelet basis and grain details)
Eh=min (max (| HL0 |)) (1)
Ev=min (max (| LH0 |)) (2)
Ep=min (max (| HL0 |)-min (| HL0 |)) (3)
Eq=min (max (| LH0 |)-min (| LH0 |)) (4)
Wherein, HLo represents the human body device of the low frequency extraction corresponding with the high fdrequency component LH1 in vertical direction in horizontal direction
Official's image, LHo represent the human organ figure of the high frequency extraction corresponding with the low frequency component HL1 in vertical direction in horizontal direction
Picture；I, j represent the sequence number of row, and i ', j ' represent the sequence number of row；S1.3-3, according to foregoing six image texture characteristic parameter acquirings
Wavelet basis filter coefficient.
The visual signature of homogeneity phenomenon in image texture reflection image, Fig. 4 clearly show texture region and express figure
The key character such as edge, graded of main target as in；We are using 6 energy and texture formula of S1.3-2 as preferred
Qualifications, so that the wave filter of selected suitable target texture image.Mainly preferred method is：According to foregoing six image textures
Characteristic parameter, wavelet basis filter coefficient is obtained using genetic programming algorithm.
In step S1.3-3 of the invention, wavelet basis filter coefficient acquisition methods are：A certain amount of sample image is extracted,
The table of comparisons of 6 characteristic indexs and wavelet basis filter coefficient is set up, directly control is carried out when calculating in real time and is searched.
As shown in figure 3, horizontal line is the calculating to row pixel, ordinate is the calculating to row pixel, and although the method is rough
Organ general image is extracted, but it is highly effective in terms of ensureing to calculate real-time, in step S1.3-1, using setting
The method extraction human organ for putting boundary threshold specifically includes following steps：
To when high fdrequency component LHn, the horizontal direction on the previous low frequency and vertical direction decomposed in the horizontal direction for obtaining
On high frequency and vertical direction on two subgraphs of low frequency component HLn carry out operations described below, respectively obtain organic image HLo
And LHo：
As sum (| Row |)>During Threshold1, retain current line pixel data, otherwise, current line pixel data is made
It is background, deletes the row data；
As sum (| Column |)>During Threshold2, retain and work as prostatitis pixel data, otherwise, prostatitis pixel data will be worked as
As background, the row data are deleted；
Wherein, Row is capable pixel, and Column is the pixel of row, and Threshold1, Threshold2 are respectively organ moulds
The corresponding parameter threshold of type.
In step S2 of the invention, two width different modalities medical images are decomposed and obtains the combination of high frequency, low frequency and low-and high-frequency
Component be overlapped under certain weight, obtain the component that the high frequency of fused images, low frequency and low-and high-frequency are combined, specially：
S2.1, the subgraph of any two mode medical image after decomposition is merged with certain weight, according to
The principle that low-and high-frequency correspondence is added is carried out, and by taking any two width different modalities medical image as an example, the fusion weight W of a sub-picture can
To select the arbitrary small number between 0.1~0.9, a secondary fusion weight is 1-W in addition；
S2.2, the decomposition result AHHn, ALLn, AHLn, ALHn and BHHn to different modalities image A and B, BLLn, BHLn,
BLHn is overlapped computing, and n represents n-th layer decomposition, if fused images are F：
FHHn=wHH*AHHn+ (1-wHH) * BHHn；WHH is that high frequency merges weight
FLLn=wLL*ALLn+ (1-wLL) * BLLn；WLL is that low frequency merges weight
FHLn=wHL*AHLn+ (1-wHL) * BHLn；WHL is that low-and high-frequency merges weight
FLHn=wLH*ALHn+ (1-wLH) * BLHn；WLH is that low high frequency merges weight.
Part that the present invention does not relate to is same as the prior art or can be realized using prior art.
Claims (9)
1. a kind of multimode medical image fusion method based on adaptive wavelet packet transform, it is characterized in that it is comprised the following steps：
S1, wavelet transformation is carried out to the medical image of different modalities using adaptive wavelet packet transform wave filter, by earlier figures picture difference
It is decomposed into the component that high frequency, low frequency and low-and high-frequency are combined；
S2, the component of the high frequency, low frequency and low-and high-frequency combination obtained to the decomposition of any two width different modalities medical image are folded
Plus, obtain the component that high frequency, low frequency and the low-and high-frequency of fused images are combined；
The component that S3, the high frequency to fused images, low frequency and low-and high-frequency are combined carries out discrete wavelet inverse transformation, obtains original size
Fused images.
2. the multimode medical image fusion method based on wavelet transformation according to claim 1, it is characterized in that described
In step S1, low frequency component is decomposed again according to yardstick demand.
3. the multimode medical image fusion method based on wavelet transformation according to claim 1, it is characterized in that described
Step S1 is specially：
S1.1, wavelet structure base filter bank, i.e., based on meeting orthogonality condition fixed length wavelet filter, in different initial values
Under the conditions of obtain multiple filter coefficients and build wavelet basis filter banks, a filtering is randomly choosed in wavelet basis filter bank
Device coefficient is used as initialization wave filter；
S1.2, wavelet transformation carried out to the medical image of multiple mode obtain ground floor decomposed component using initialization wave filter,
I.e. by wavelet transform obtain low frequency in image low frequency component LL1 in the horizontal and vertical directions, horizontal direction and
Low frequency component HL1 on high frequency and vertical direction on high fdrequency component LH1, horizontal direction and level in vertical direction and hang down
The upward high fdrequency component HH1 of Nogata；
The fluctuation of S1.3, the Energy distribution according to image and texture selects filter coefficient to step again in wavelet basis filter bank
Low frequency component LL1 in the ground floor decomposed component that rapid S1.2 is obtained carries out second wavelet transformation, obtain LL2, HH2, HL2,
LH2 repeat steps S1.3 several times, completes picture breakdown.
4. the multimode medical image fusion method based on wavelet transformation according to claim 3, it is characterized in that described
In step S1.3, the number of times that repeat step S1.3 carries out picture breakdown is 3-5 times.
5. the multimode medical image fusion method based on wavelet transformation according to claim 3, it is characterized in that described
In step S1.3, Energy distribution and texture according to image fluctuate and the specific of filter coefficient are selected in wavelet basis filter bank
Method is：
S1.3-1, the step of carry out organ to image and extract：
Four components in current decomposition component are processed, is extracted human organ from the background of black, specifically
Using image partition method extract human organ, or using set boundary threshold method extract human organ image HLo and
LHo；
The step of S1.3-2, calculating image texture characteristic parameter：
It is minimum horizontal image least energy Eh, the image least energy Ev of longitudinal direction, horizontal image to be calculated using following formula
Extreme difference Ep, image minimum extreme difference Eq, the transverse direction lines maximum fluctuation value Em and longitudinal lines maximum fluctuation value En of longitudinal direction；
Eh=min (max (| HLo |)) (1)
Ev=min (max (| LHo |)) (2)
Ep=min (max (| HLo |)-min (| HLo |)) (3)
Eq=min (max (| LHo |)-min (| LHo |)) (4)
Wherein, HLo represents the human organ figure of the low frequency extraction corresponding with the high fdrequency component LH1 in vertical direction in horizontal direction
Picture, LHo represent the human organ image of the high frequency extraction corresponding with the low frequency component HL1 in vertical direction in horizontal direction；i、j
The sequence number of row is represented, i ', j ' represent the sequence number of row；S1.3-3, according to foregoing six image texture characteristics parameter acquiring wavelet basis
Filter coefficient.
6. the multimode medical image fusion method based on wavelet transformation according to claim 5, it is characterized in that described
In step S1.3-3, wavelet basis filter coefficient acquisition methods are：According to six image texture characteristics that step S1.3-2 is obtained
Parameter, wavelet basis filter coefficient, preferably genetic programming algorithm are obtained using optimization algorithm.
7. the multimode medical image fusion method based on wavelet transformation according to claim 5, it is characterized in that described
In step S1.3-3, wavelet basis filter coefficient acquisition methods are：A certain amount of sample image is extracted, six characteristic indexs are set up
With the table of comparisons of wavelet basis filter coefficient, in real time calculate when, according to step S1.3-2 obtain six image texture characteristics
Parameter control is searched and obtains corresponding wavelet basis filter coefficient.
8. the multimode medical image fusion method based on wavelet transformation according to claim 5, it is characterized in that described
In step S1.3-1, human organ is extracted using the method for setting boundary threshold and specifically includes following steps：
To as the high fdrequency component LHn on the previous low frequency and vertical direction decomposed in the horizontal direction for obtaining, in horizontal direction
Two subgraphs of the low frequency component HLn on high frequency and vertical direction carry out operations described below, respectively obtain organic image HLo and
LHo：
As sum (| Row |)>During Threshold1, retain current line pixel data, otherwise, using current line pixel data as the back of the body
Scape, deletes the row data；
As sum (| Column |)>During Threshold2, retain and work as prostatitis pixel data, otherwise, prostatitis pixel data conduct will be worked as
Background, deletes the row data；
Wherein, Row is capable pixel, and Column is the pixel of row, and Threshold1, Threshold2 are respectively organ models pair
The parameter threshold answered.
9. the multimode medical image fusion method based on wavelet transformation according to claim 1, it is characterized in that described
In step S2, two width different modalities medical images are decomposed the component for obtaining the combination of high frequency, low frequency and low-and high-frequency in certain weight
Under be overlapped, obtain the component that the high frequency of fused images, low frequency and low-and high-frequency are combined, specially：
S2.1, the subgraph of any two mode medical image after decomposition is merged with certain weight, according to height
The principle that frequency correspondence is added is carried out, wherein the fusion weight W of a sub-picture selects the arbitrary small number between 0.1~0.9, in addition one
Pair fusion weight is 1-W；
S2.2, the decomposition result AHHn, ALLn, AHLn, ALHn and BHHn to different modalities image A and B, BLLn, BHLn, BLHn
Computing is overlapped, n represents n-th layer decomposition, obtains four component Fs HHn, FLLn, FHLn and the FLHn of fused images F：
FHHn=wHH*AHHn+ (1-wHH) * BHHn；WHH is that high frequency merges weight
FLLn=wLL*ALLn+ (1-wLL) * BLLn；WLL is that low frequency merges weight
FHLn=wHL*AHLn+ (1-wHL) * BHLn；WHL is that low-and high-frequency merges weight
FLHn=wLH*ALHn+ (1-wLH) * BLHn；WLH is that low high frequency merges weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710046867.4A CN106910179B (en) | 2017-01-22 | 2017-01-22 | Multimode medical image fusion method based on wavelet transformation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710046867.4A CN106910179B (en) | 2017-01-22 | 2017-01-22 | Multimode medical image fusion method based on wavelet transformation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106910179A true CN106910179A (en) | 2017-06-30 |
| CN106910179B CN106910179B (en) | 2019-03-15 |
Family
ID=59207184
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710046867.4A Active CN106910179B (en) | 2017-01-22 | 2017-01-22 | Multimode medical image fusion method based on wavelet transformation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106910179B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107742279A (en) * | 2017-10-31 | 2018-02-27 | 努比亚技术有限公司 | A kind of image processing method, device and storage medium |
| CN107993236A (en) * | 2017-11-27 | 2018-05-04 | 上海交通大学 | A kind of method and platform of multi-modality images processing |
| CN109544550A (en) * | 2018-12-05 | 2019-03-29 | 易必祥 | Based on CT images intelligent measurement and recognition methods and system |
| CN109657709A (en) * | 2018-12-06 | 2019-04-19 | 湖北博华自动化系统工程有限公司 | A kind of equipment fault prediction technique based on particle group optimizing support vector regression |
| CN109919929A (en) * | 2019-03-06 | 2019-06-21 | 电子科技大学 | A kind of fissuring of tongue feature extracting method based on wavelet transformation |
| CN111445492A (en) * | 2020-03-26 | 2020-07-24 | 北京易康医疗科技有限公司 | Three-dimensional fusion calibration method for magnetic resonance image and radiotherapy positioning image |
| CN111667486A (en) * | 2020-04-29 | 2020-09-15 | 杭州深睿博联科技有限公司 | Multi-mode fusion pancreas segmentation method and system based on deep learning |
| CN112932497A (en) * | 2021-03-10 | 2021-06-11 | 中山大学 | Unbalanced single-lead electrocardiogram data classification method and system |
| CN118037560A (en) * | 2024-01-16 | 2024-05-14 | 北京长木谷医疗科技股份有限公司 | Homomorphic filtering-based multi-mode medical image fusion method, device and equipment |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105844606A (en) * | 2016-03-22 | 2016-08-10 | 博康智能网络科技股份有限公司 | Wavelet transform-based image fusion method and system thereof |
-
2017
- 2017-01-22 CN CN201710046867.4A patent/CN106910179B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105844606A (en) * | 2016-03-22 | 2016-08-10 | 博康智能网络科技股份有限公司 | Wavelet transform-based image fusion method and system thereof |
Non-Patent Citations (3)
| Title |
|---|
| 张彬 等: "基于小波变换的多模态医学图像融合算法", 《重庆医学》 * |
| 徐磊 等: "基于离散小波变换的多模态医学图像融合改进算法的研究", 《重庆医学》 * |
| 詹玲超等: "基于小波变换的多模态医学图像的融合方法", 《现代电子技术》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107742279B (en) * | 2017-10-31 | 2020-07-10 | 珠海大横琴科技发展有限公司 | Image processing method, device and storage medium |
| CN107742279A (en) * | 2017-10-31 | 2018-02-27 | 努比亚技术有限公司 | A kind of image processing method, device and storage medium |
| CN107993236A (en) * | 2017-11-27 | 2018-05-04 | 上海交通大学 | A kind of method and platform of multi-modality images processing |
| CN109544550A (en) * | 2018-12-05 | 2019-03-29 | 易必祥 | Based on CT images intelligent measurement and recognition methods and system |
| CN109544550B (en) * | 2018-12-05 | 2021-10-22 | 易必祥 | CT image-based intelligent detection and identification method and system |
| CN109657709A (en) * | 2018-12-06 | 2019-04-19 | 湖北博华自动化系统工程有限公司 | A kind of equipment fault prediction technique based on particle group optimizing support vector regression |
| CN109919929A (en) * | 2019-03-06 | 2019-06-21 | 电子科技大学 | A kind of fissuring of tongue feature extracting method based on wavelet transformation |
| CN109919929B (en) * | 2019-03-06 | 2023-03-24 | 电子科技大学 | Tongue crack feature extraction method based on wavelet transformation |
| CN111445492A (en) * | 2020-03-26 | 2020-07-24 | 北京易康医疗科技有限公司 | Three-dimensional fusion calibration method for magnetic resonance image and radiotherapy positioning image |
| CN111667486A (en) * | 2020-04-29 | 2020-09-15 | 杭州深睿博联科技有限公司 | Multi-mode fusion pancreas segmentation method and system based on deep learning |
| CN111667486B (en) * | 2020-04-29 | 2023-11-17 | 杭州深睿博联科技有限公司 | Multi-modal fusion pancreas segmentation method and system based on deep learning |
| CN112932497A (en) * | 2021-03-10 | 2021-06-11 | 中山大学 | Unbalanced single-lead electrocardiogram data classification method and system |
| CN118037560A (en) * | 2024-01-16 | 2024-05-14 | 北京长木谷医疗科技股份有限公司 | Homomorphic filtering-based multi-mode medical image fusion method, device and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106910179B (en) | 2019-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106910179B (en) | Multimode medical image fusion method based on wavelet transformation | |
| Yang et al. | Low-dose CT image denoising using a generative adversarial network with Wasserstein distance and perceptual loss | |
| CN104809734B (en) | A method of the infrared image based on guiding filtering and visual image fusion | |
| CN103985099B (en) | Dispersion tensor magnetic resonance image tensor domain non-local mean denoising method | |
| CN100561518C (en) | Self-adaptation medical image sequence interpolation method based on area-of-interest | |
| KR20190104036A (en) | Treat patients with optimized tumor therapies with templates with variable electrode positions | |
| CN109934887B (en) | Medical image fusion method based on improved pulse coupling neural network | |
| Gupta | Nonsubsampled shearlet domain fusion techniques for CT–MR neurological images using improved biological inspired neural model | |
| CN103501699A (en) | Method and apparatus for isolating a potential anomaly in imaging data and its application to medical imagery | |
| CN110660063A (en) | Multi-image fused tumor three-dimensional position accurate positioning system | |
| CN107507189A (en) | Mouse CT image kidney dividing methods based on random forest and statistical model | |
| CN106504221B (en) | Method of Medical Image Fusion based on quaternion wavelet transformation context mechanism | |
| CN103617604B (en) | Fusion method based on the image that two-dimensional empirical mode decomposition method feature extracts | |
| Yang | Multiresolution Image Fusion Based on Wavelet Transform By Using a Novel Technique for Selection Coefficients. | |
| CN102682439B (en) | Medical image fusion method based on multidirectional empirical mode decomposition | |
| Dogra et al. | Multi-modality medical image fusion based on guided filter and image statistics in multidirectional shearlet transform domain | |
| Zhou et al. | N-net: 3d fully convolution network-based vertebrae segmentation from ct spinal images | |
| Patel et al. | Medical image fusion based on multi-scaling (DRT) and multi-resolution (DWT) technique | |
| Rousseau et al. | A supervised patch-based image reconstruction technique: Application to brain MRI super-resolution | |
| Zhang et al. | Research on image fusion algorithm based on NSST frequency division and improved LSCN | |
| Zhang et al. | Medical image fusion based a densely connected convolutional networks | |
| CN115018728A (en) | Image fusion method and system based on multi-scale transformation and convolution sparse representation | |
| CN115619693A (en) | Space-time fusion model based on nested multi-scale transformation | |
| Yang et al. | Medical image fusion method based on lifting wavelet transform and dual-channel PCNN | |
| CN110084772B (en) | MRI/CT fusion method based on bending wave |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Shiping Inventor after: Wang Xiaofang Inventor before: Wang Xiaofang |
|
| CB03 | Change of inventor or designer information | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20181026 Address after: 518000 B1 701-72, Kexing Science Park, 15 Keyuan Road, Nanshan District, Shenzhen, Guangdong. Applicant after: Cabin (Shenzhen) Medical Technology Co.,Ltd. Address before: 210000 Room 403, 3 tower 141, Ma Tai Street, Nanjing, Jiangsu. Applicant before: NANJING MIZONG ELECTRONIC TECHNOLOGY Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200525 Address after: 518101 B05, floor 3, workshop 1, Daqian industrial plant, zone 67, Xingdong community, Xin'an street, Bao'an District, Shenzhen City, Guangdong Province Patentee after: Kaben (Shenzhen) medical equipment Co.,Ltd. Address before: 518000 B1 701-72, Kexing Science Park, 15 Keyuan Road, Nanshan District, Shenzhen, Guangdong. Patentee before: Cabin (Shenzhen) Medical Technology Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Multimodal medical image fusion method based on Wavelet Transform Effective date of registration: 20211229 Granted publication date: 20190315 Pledgee: Shenzhen hi tech investment small loan Co.,Ltd. Pledgor: Kaben (Shenzhen) medical equipment Co.,Ltd. Registration number: Y2021980017017 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PC01 | Cancellation of the registration of the contract for pledge of patent right |
Date of cancellation: 20230410 Granted publication date: 20190315 Pledgee: Shenzhen hi tech investment small loan Co.,Ltd. Pledgor: Kaben (Shenzhen) medical equipment Co.,Ltd. Registration number: Y2021980017017 |
|
| PC01 | Cancellation of the registration of the contract for pledge of patent right |