CN102592137A - Multi-modality image registration method and operation navigation method based on multi-modality image registration - Google Patents
Multi-modality image registration method and operation navigation method based on multi-modality image registration Download PDFInfo
- Publication number
- CN102592137A CN102592137A CN2011104447720A CN201110444772A CN102592137A CN 102592137 A CN102592137 A CN 102592137A CN 2011104447720 A CN2011104447720 A CN 2011104447720A CN 201110444772 A CN201110444772 A CN 201110444772A CN 102592137 A CN102592137 A CN 102592137A
- Authority
- CN
- China
- Prior art keywords
- registration
- image
- art
- sampling
- magnetic resonance
- 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
Images
Abstract
A multi-modality image registration method includes the following steps: sampling reference images and images to be registered according to a first sampling proportion to obtain the reference images and the images to be registered with first resolution ratio, adopting a B spline surface as a deformation model to perform matching on the reference images and the images to be registered, calculating matching measure and obtaining first matching parameters, and increasing sampling proportion to repeat the processes until the matching measure reaches a reset threshold value, wherein in matching, matching parameters of last time are used as initial matching parameters after each-time sampling proportion is increased. An operation navigation method based on multi-modality image registration is further provided.
Description
[technical field]
The present invention relates to Flame Image Process, relate in particular to a kind of multi-modality images method for registering and reach surgical navigational method based on the multi-modality images registration.
[background technology]
In the prior art, domestic interventional navigation system is primarily aimed at bone and the operation of brain section, and how based on the registration of ultrasonic image.Because the rigidity characteristics of bone make the registration between preceding image of art and the patient accomplish Rigid Registration than being easier to through the monumented point on the bone, also can reach reasonable treatment precision.But for the operation of soft tissues such as liver, because the effect of breathing or operating theater instruments, unavoidable ground can cause the displacement and the distortion of liver.At this moment, just no longer suitable based on the Rigid Registration of unique point.And general ultrasonic image noise is big and blur margin is clear, because the imaging technique of ultrasonoscopy limitation still is in conceptual phase based on ultrasonic high precision image method for registering.And registration speed is technological emphasis, difficult point to the elastic registrating technology of yielding tissues such as liver.
Therefore, present interventional navigation system in the real-time of registration, all shortcomings to some extent of aspects such as three-dimensional image information, operation pathway planning and navigation accuracy are provided.Because different image documentation equipment adopts different formation method images acquired, every kind of image-forming principle can only reflect the information of certain organ effectively, and for example CT can clear demonstration skeleton, and MRI equipment then helps showing soft tissue.Also there is not a kind of imaging technique can obtain human organ information all sidedly at present, and fast and accurately with the image registration of various organ information.
[summary of the invention]
Based on this, be necessary to provide a kind of method for registering of multi-modality images fast and accurately.
A kind of multi-modality images method for registering may further comprise the steps:
With reference picture with treat that registering images is sampled according to the first sampling ratio and obtain reference picture and treat registering images with first resolution;
Adopt B-spline surface as distorted pattern to said reference picture with first resolution with treat that registering images carries out registration, calculates registration and estimate and obtain first registration parameter;
Increasing the sampling ratio repeats above-mentioned steps and estimates until registration and reach predetermined threshold value; Wherein, the registration parameter that at every turn increases when carrying out registration after the sampling ratio all with the last time is the initial registration parameter.
Preferably, said to reference picture with treat that the step that registering images is sampled according to the sampling ratio specifically comprises:
With said reference picture with treat that the grey scale pixel value of registering images carries out normalization;
With the sampling ratio 256 gray levels are sampled.
Preferably, the sampling and the step of registration repeat 1 time, wherein, and according to reference picture and treat that the resolution of registering images is provided with the first sampling ratio and repeated sampling ratio.
Preferably, the method for sampling adopts one of following mode: stochastic sampling, equal interval sampling and pyramid stratified sampling.
Preferably, adopt gradient descent algorithm, step-size in search, the relaxation factor in the condition of convergence, maximum iteration time, the B-spline surface size of first sampling and repeated sampling is set to the registration optimized Algorithm.
Preferably, in the step that said calculating registration is estimated:
Adopt normalized mutual information value to estimate as registration.
Preferably, registration is estimated when reaching predetermined threshold value, with said reference picture with treat that registering images carries out absolute registration, definition B batten control mesh in absolute registration wherein.
In addition, also be necessary to provide a kind of fast and accurately based on the surgical navigational method of multi-modality images registration.
A kind of surgical navigational method based on the multi-modality images registration, the 3-D view that is used for setting up according to CT image before the art and the nuclear magnetic resonance image of art carry out the surgical navigational behind the registration, comprise the steps:
The CT image is rebuild and is obtained DRR (digitally reconstructed radiography, digital reconstruction image) image before adopting the digitized video reconstruction technique to art;
Nuclear magnetic resonance image in DRR image and the art is carried out registration and constantly changes the space position parameter of DRR image in simulation x-ray imaging equipment, obtain the asynchronous registration of space position parameter and estimate;
Search out the space position parameter that registration estimates when arriving threshold value through optimized Algorithm and be the space orientation parameter of DRR image at the simulation X ray;
Euler transformation parameter when obtaining the nuclear magnetic resonance image similarity maximum in DRR image and the art is as the coordinate transformation parameter that the CT view data before the art is transformed into nuclear magnetic resonance image;
Combine before the said art navigation that undergos surgery of the nuclear magnetic resonance image in the CT image and art according to said coordinate transformation parameter;
Wherein, the said step that nuclear magnetic resonance image in DRR image and the art is carried out registration comprises:
Nuclear magnetic resonance image in DRR image and the art sampled according to the first sampling ratio obtain DRR image with first resolution and the nuclear magnetic resonance image in the art;
Adopt B-spline surface as distorted pattern to said DRR image with first resolution carry out registration with the nuclear magnetic resonance image in the art, the calculating registration is estimated and obtain first registration parameter;
Increasing the sampling ratio repeats above-mentioned step of registration and estimates until registration and reach predetermined threshold value; Wherein, the registration parameter that at every turn increases when carrying out registration after the sampling ratio all with the last time is the initial registration parameter.
Preferably, said nuclear magnetic resonance image in DRR image and the art is sampled according to the first sampling ratio obtained DRR image with first resolution and the step of the nuclear magnetic resonance image in the art specifically comprises:
The grey scale pixel value of the nuclear magnetic resonance image in said DRR image and the art is carried out normalization;
With the said first sampling ratio 256 gray levels are sampled.
Preferably, the sampling and the step of registration repeat 1 time, wherein, and according to reference picture and treat that the resolution of registering images is provided with the first sampling ratio and repeated sampling ratio.
Preferably, adopt gradient descent algorithm, step-size in search, the relaxation factor in the condition of convergence, maximum iteration time, the B-spline surface size of first sampling and repeated sampling is set to the registration optimized Algorithm.
Preferably, in the step that said calculating registration is estimated:
Adopt normalized mutual information value to estimate as registration.
Preferably, registration is estimated when reaching predetermined threshold value, and the nuclear magnetic resonance image in said DRR image and the art is carried out absolute registration, wherein definition B batten control mesh in absolute registration.
Preferably, saidly combine the undergo surgery step of navigation of the nuclear magnetic resonance image in the CT image and art before the said art to comprise according to said coordinate transformation parameter:
CT image before the art that similarity is maximum and the nuclear magnetic resonance image in the art carry out image co-registration according to the space matching relationship, and according to the Image Acquisition surgical navigational path after the said fusion.
Above-mentioned multi-modality images method for registering; Adopt multilayer image in different resolution registration implementation method; Simultaneously with B-spline surface as distorted pattern, can be fast and accurately with reference picture with treat that registering images finds both registrations to estimate threshold value, realizes image registration accurately.
Above-mentioned surgical navigational method based on the multi-modality images registration; Adopt multilayer image in different resolution registration implementation method; Simultaneously with B-spline surface as distorted pattern, can be fast and accurately with reference picture with treat that registering images finds both registrations to estimate threshold value, realizes image registration accurately; And the image behind the registration merged, can obtain that the multi-source channel gathers to about the favourable information in each channel of the data of same target.
[description of drawings]
Fig. 1 is the process flow diagram of multi-modality images method for registering;
Fig. 2 is for extracting the synoptic diagram of area-of-interest;
Fig. 3 is the synoptic diagram that dwindles B batten grid control area;
Fig. 4 (a) to Fig. 4 (c) be multi-modality images method for registering implementation procedure synoptic diagram;
Fig. 5 is the process flow diagram based on the surgical navigational method of multi-modality images registration.
[embodiment]
Because multi-modality images is the image of taking under different imaging devices and the different shooting times.Therefore reference picture be in the art preplanning system target image with treat that registering images is a image in the art and inequality, for example, both are not at same one deck, perhaps two kinds of images have distortion or offset, therefore need to proofread and correct in good time.Therefore the present invention adopts following multi-modality images method for registering.
As shown in Figure 1, the process flow diagram for a kind of multi-modality images method for registering may further comprise the steps:
Step S110, with reference picture with treat that registering images is sampled according to the first sampling ratio and obtain reference picture and treat registering images with first resolution.
In the present embodiment, among the step S110 to reference picture with treat that the step that registering images is sampled according to the sampling ratio specifically comprises:
1. with said reference picture with treat that the grey scale pixel value of registering images carries out normalization.
2. with the sampling ratio 256 gray levels are sampled.
In the present embodiment, the method for sampling adopts one of following mode: stochastic sampling, equal interval sampling and pyramid stratified sampling.Preferably, adopt the pyramid stratified sampling.
Image registration is the process to the multiple image coupling of the Same Scene of taking from different time, different sensors or different visual angles.In general, the original image resolution of CT image and MRI can reach various high resolving power.Therefore the grey scale pixel value with image carries out normalization, with the data sampling of certain proportion to whole 256 gray levels, and adopts sampling subset to replace all pixel datas to calculate similarity measures, and promptly each time registration adopts pictures different resolution.
In the present embodiment; All adopt 3 B-spline surfaces as the space conversion function each the employing in the different resolution registration process; And, can in every layer registration process, adopt the B-spline surface that varies in size as distorted pattern according to the registration needs of different scale.Preferably, the true resolution according to multi-modality images adopts the B-spline surface that varies in size as distorted pattern.
Step S120, adopt B-spline surface as distorted pattern to said reference picture with first resolution with treat that registering images carries out registration, calculates registration and estimate and obtain first registration parameter.
In the present embodiment, in the step that the calculating registration is estimated: adopt normalized mutual information value to estimate as registration.
Mutual information is a key concept in the information theory, is used to describe two statistic correlations between system, or the amount of information in another system that is comprised in system.Wherein, in registration is estimated, because variance and mean square deviation are to estimate most important, the most frequently used index of data variation degree, the normalized form of therefore normal employing.
In the present embodiment, adopt gradient descent algorithm, step-size in search, the relaxation factor in the condition of convergence, maximum iteration time, the B-spline surface size of first sampling and repeated sampling is set to the registration optimized Algorithm.
In the present embodiment, adopt gradient descent algorithm to the registration optimized Algorithm of first sampling, wherein step-size in search is that the relaxation factor in 10mm, the condition of convergence is 0.7, maximum iteration time is 200, adopts 5 * 5 B-spline surface as deformation model; Adopt the LBFG algorithm to the registration optimized Algorithm of repeated sampling, wherein step-size in search is that 0.1mm, gradient convergence tolerance are 0.05, and maximum iteration time is 1000 times, adopt 10 * 10 B-spline surface as deformation model.First registration parameter comprises: pixel, contraction coefficient and amplitude length etc.
Step S130 increases the sampling ratio and repeats above-mentioned steps and estimate until registration and reach predetermined threshold value; Wherein, the registration parameter that at every turn increases when carrying out registration after the sampling ratio all with the last time is the initial registration parameter.
The registration parameter of last time as next time initial registration parameter, can be improved the continuity and the stability of twice registration.
In the present embodiment, the sampling and the step of registration repeat 1 time, wherein, and according to reference picture and treat that the resolution of registering images is provided with the first sampling ratio and repeated sampling ratio.
In the present embodiment, the step of sampling and registration repeats 1 time, and wherein, the first sampling ratio is 0.3, and the repeated sampling ratio is 0.6.
Show like table 1, be image registration results, comprise the contrast of the mutual information of registration result and reference picture after mutual information value, first registration result and the second registration result registration before the registration accomplished.
| The mutual information value | Before the registration | First registration result | Second registration result |
| 1 | 1.24257 | 1.27905 | 1.28651 |
Table 1
As shown in table 2 is the registration result of CT (512*512) and MRI (512*416) two width of cloth different modalities images, and wherein the maximum mutual information value is as similarity measure.The registration output parameter comprises preceding reference diagram of registration and the mutual information value of floating and scheming, the contrast of mutual information value and the registration required time of first registration result and second registration result and reference diagram.It is more to know that the mutual information value promotes the mutual information value contrast before and after the registration, explain that floating image and the consistance of reference picture after being out of shape increases, and this method for registering has certain registration effect for multi-modality images.
| The mutual information value | Before the registration | First registration result | Second registration result | The registration time |
| 1 | 0.666575 | 0.924476 | 0.924923 | 208 |
Table 2
The result of experiment that table 3 designs for the registration speed of registration strategies in the present embodiment relatively and conventional monolayers method for registering.Traditional single method for registering in this experiment adopts identical method for registering, comprises using B-spline surface to combine the LBFG optimized Algorithm to seek the registration strategies of optimal value as distorted pattern, mutual information as similarity measure.On this basis, can think that two kinds of method registration conditions of convergence are identical, promptly registration accuracy is suitable.When the same registration condition of convergence was set, the multilayer method for registering registration required time of single method for registering and this paper was as shown in the table:
| Time (S) CT/MRI | Classic method | Multiresolution |
| 1 | 451 | 208 |
Table 3
Registration is estimated when reaching predetermined threshold value, with reference picture with treat that registering images carries out absolute registration, definition B batten control mesh in absolute registration wherein.When reference picture and when treating that difference between the registration parameter of registering images is between ± 5%, think that registration is estimated and reach predetermined threshold value.Particularly, adopt the method step of B batten grid control area to comprise:
1. choose the area-of-interest in the reference picture, obtain its Pixel Information, and Pixel Information is mapped to the corresponding region of treating registering images, control B batten grid simultaneously and change.
Be illustrated in figure 2 as the synoptic diagram that extracts area-of-interest.Area-of-interest 10 is the deformed region in the reference picture, area-of-interest 10 is extracted, and be mapped to the corresponding region of treating registering images, and control B batten grid changes.Preferably, in B batten grid, area-of-interest 10 is dwindled.
2. the grid control area after will changing is optimized calculating, reaches to be embedded into after the matching threshold to treat in the registering images.
As shown in Figure 3, for dwindling the synoptic diagram of B batten grid control area.Deformed region 20 is the moderate finite deformation zones in the reference picture, dwindles zone 22 and is the grid control area after changing.22 be optimized calculating to dwindling the zone, reach to be embedded into after the matching threshold and treat in the registering images.Concrete implementation procedure is as shown in Figure 4: Fig. 4 a is a reference picture.Fig. 4 b is for treating registering images.Fig. 4 c is a registration result.
Next be the multilayer registration strategies of contrast multi-modality images and the registration speed of conventional monolayers method for registering, experiment below the design.The identical registration parameter of conventional monolayers method for registering and layering method for registering employing in this experiment comprises that same use B-spline surface estimates the registration strategies of seeking optimal value with the LBFG optimized Algorithm as distorted pattern, mutual information as registration.On this basis, can think that two kinds of method registration conditions of convergence are identical, promptly registration accuracy is suitable.When the same registration condition of convergence was set, the multilayer method for registering registration required time of individual layer method for registering and multi-modality images was as shown in table 4.
| Multi-modality images | The classic method registration time (s) | The layering registration time (s) |
| 1 | 451 | 208 |
| 2 | 100 | 28 |
| 3 | 150 | 15 |
Table 4
At B batten number of times, the constant situation of control B batten mesh spacing, said method has improved speed and the precision calculated, is more suitable for revising the image of local moderate finite deformation.
Based on above-mentioned all embodiment; As shown in Figure 5; Be a kind of surgical navigational method flow diagram based on the multi-modality images registration, the 3-D view that is used for setting up according to CT image before the art and the nuclear magnetic resonance image of art carry out the surgical navigational behind the registration, comprise the steps:
Step S210, adopt the digitized video reconstruction technique to art before the CT image rebuild and obtain the DRR image.
Adopt digitized video reconstruction technique (DRR) to the CT image reconstruction before the art, promptly adopt the preceding CT layer data of art to reconstruct affine photographs, the physical process of simulation actual imaging.
In the present embodiment, the CT image before the art being rebuild, is in order to realize that three-dimensional art preplanning system is reconstructed into three-dimensional with the CT sectioning image.
Step S220 carries out the nuclear magnetic resonance image in DRR image and the art registration and constantly changes the space position parameter of DRR image in simulation x-ray imaging equipment, obtains the asynchronous registration of space position parameter and estimates.
In the present embodiment, therefore the parameter of six-freedom degrees such as the rotation of employing Euler Spin Control 3-D view and translation in the process that resamples, changes the euler transformation parameter, can obtain the DRR image of different angles.
The step of in the present embodiment, the nuclear magnetic resonance image in DRR image and the art being carried out registration comprises:
1. the nuclear magnetic resonance image in DRR image and the art is sampled according to the first sampling ratio and obtain DRR image with first resolution and the nuclear magnetic resonance image in the art.
2. adopt B-spline surface said DRR image with first resolution and the nuclear magnetic resonance image in the art to be carried out registration, calculate registration and estimate and obtain first registration parameter as distorted pattern.
3. increasing the sampling ratio repeats above-mentioned step of registration and estimates until registration and reach predetermined threshold value; Wherein, the registration parameter that at every turn increases when carrying out registration after the sampling ratio all with the last time is the initial registration parameter.
In the present embodiment, the nuclear magnetic resonance image in DRR image and the art is compared according to first sampling
Example is sampled and is obtained DRR image with first resolution and the step of the nuclear magnetic resonance image in the art specifically comprises:
1. the grey scale pixel value with the nuclear magnetic resonance image in said DRR image and the art carries out normalization.
2. with the sampling ratio 256 gray levels are sampled.
In the present embodiment, the sampling and the step of registration repeat 1 time, wherein, and according to reference picture and treat that the resolution of registering images is provided with the first sampling ratio and repeated sampling ratio.
In the present embodiment, the step of sampling and registration repeats 1 time, and wherein, the first sampling ratio is 0.3, and the repeated sampling ratio is 0.6.
Resampling commonly used has arest neighbors method, bilinear interpolation method and three convolution interpolating methods.Wherein, the arest neighbors method is the simplest, and computing velocity is fast, but visual effect is poor; Bilinear interpolation can make image outline fuzzy; The image of three convolution method generations is more level and smooth, good visual effect is arranged, but calculated amount is big, and is more time-consuming.In the present embodiment, the image for the visual effect that obtains generally can adopt three convolution methods to resample.
In the present embodiment, adopt gradient descent algorithm, step-size in search, the relaxation factor in the condition of convergence, maximum iteration time, the B-spline surface size of first sampling and repeated sampling is set to the registration optimized Algorithm.
In the present embodiment, adopt gradient descent algorithm to the registration optimized Algorithm of first sampling, wherein step-size in search is that the relaxation factor in 10mm, the condition of convergence is 0.7, maximum iteration time is 200, adopts 5 * 5 B-spline surface as deformation model; Registration optimized Algorithm during to repeated sampling adopts the LBFG algorithm, and wherein step-size in search is that 0.1mm, gradient convergence tolerance are 0.05, and maximum iteration time is 1000 times, adopt 10 * 10 B-spline surface as deformation model.
In the present embodiment, registration is estimated when reaching predetermined threshold value, will carry out absolute registration according to 3-D view and the nuclear magnetic resonance image in the art that CT image before the art is set up, wherein definition B batten control mesh in absolute registration.
Step S230 searches out the space position parameter that registration estimates when arriving threshold value through optimized Algorithm and is the space orientation parameter of DRR image at the simulation X ray.
Step S240, the euler transformation parameter when getting the nuclear magnetic resonance image similarity maximum in DRR image and the art is as the coordinate transformation parameter that the CT view data before the art is transformed into nuclear magnetic resonance image.
In the present embodiment, when the nuclear magnetic resonance image similarity in DRR image and the art is maximum, can think the parameter of euler transformation corresponding be exactly the coordinate conversion matrix of the preceding CT image of nuclear magnetic resonance image and art in the art.
Step S250 combines before the said art navigation that undergos surgery of the nuclear magnetic resonance image in the CT image and art according to said coordinate transformation parameter.
In the present embodiment, the 3-D view of setting up according to CT image before the art and the nuclear magnetic resonance image in the art according to the fusion of space matching relationship picture after, can obtain to show the image of bone and soft tissue comprehensively.Behind the comprehensive image of acquired information, guide the control mechanical arm to accomplish the insertion type treatment according to the volume coordinate relation, accomplish the treatment navigation procedure.
Said method can provide auxiliary completion of art preplanning, teaching and training and mechanical arm to treat.
The above embodiment has only expressed several kinds of embodiments of the present invention, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to claim of the present invention.Should be pointed out that for the person of ordinary skill of the art under the prerequisite that does not break away from the present invention's design, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with accompanying claims.
Claims (14)
1. multi-modality images method for registering may further comprise the steps:
With reference picture with treat that registering images is sampled according to the first sampling ratio and obtain reference picture and treat registering images with first resolution;
Adopt B-spline surface as distorted pattern to said reference picture with first resolution with treat that registering images carries out registration, calculates registration and estimate and obtain first registration parameter;
Increasing the sampling ratio repeats above-mentioned steps and estimates until registration and reach predetermined threshold value; Wherein, the registration parameter that at every turn increases when carrying out registration after the sampling ratio all with the last time is the initial registration parameter.
2. multi-modality images method for registering according to claim 1 is characterized in that, said to reference picture with treat that the step that registering images is sampled according to the sampling ratio specifically comprises:
With said reference picture with treat that the grey scale pixel value of registering images carries out normalization;
With the sampling ratio 256 gray levels are sampled.
3. multi-modality images method for registering according to claim 1 is characterized in that, the sampling and the step of registration repeat 1 time, wherein, and according to reference picture and treat that the resolution of registering images is provided with the first sampling ratio and repeated sampling ratio.
4. according to any described multi-modality images method for registering of claim 1 to 3, it is characterized in that the method for sampling adopts one of following mode: stochastic sampling, equal interval sampling and pyramid stratified sampling.
5. multi-modality images method for registering according to claim 3; It is characterized in that; Adopt gradient descent algorithm to the registration optimized Algorithm, step-size in search, the relaxation factor in the condition of convergence, maximum iteration time, the B-spline surface size of first sampling and repeated sampling is set.
6. multi-modality images method for registering according to claim 1 is characterized in that, in the step that said calculating registration is estimated:
Adopt normalized mutual information value to estimate as registration.
7. multi-modality images method for registering according to claim 1 is characterized in that registration is estimated when reaching predetermined threshold value, with said reference picture with treat that registering images carries out absolute registration, definition B batten control mesh in absolute registration wherein.
8. surgical navigational method based on the multi-modality images registration, the 3-D view that is used for setting up according to CT image before the art and the nuclear magnetic resonance image of art carry out the surgical navigational behind the registration, comprise the steps:
The CT image is rebuild and is obtained the DRR image before adopting the digitized video reconstruction technique to art;
Nuclear magnetic resonance image in DRR image and the art is carried out registration and constantly changes the space position parameter of DRR image in simulation x-ray imaging equipment, obtain the asynchronous registration of space position parameter and estimate;
Search out the space position parameter that registration estimates when arriving threshold value through optimized Algorithm and be the space orientation parameter of DRR image in the simulation X ray;
Euler transformation parameter when getting the nuclear magnetic resonance image similarity maximum in DRR image and the art is as the coordinate transformation parameter that the CT view data before the art is transformed into nuclear magnetic resonance image;
Combine before the said art navigation that undergos surgery of the nuclear magnetic resonance image in the CT image and art according to said coordinate transformation parameter;
Wherein, the said step that nuclear magnetic resonance image in DRR image and the art is carried out registration comprises:
Nuclear magnetic resonance image in DRR image and the art sampled according to the first sampling ratio obtain DRR image with first resolution and the nuclear magnetic resonance image in the art;
Adopt B-spline surface as distorted pattern to said DRR image with first resolution carry out registration with the nuclear magnetic resonance image in the art, the calculating registration is estimated and obtain first registration parameter;
Increasing the sampling ratio repeats above-mentioned step of registration and estimates until registration and reach predetermined threshold value; Wherein, the registration parameter that at every turn increases when carrying out registration after the sampling ratio all with the last time is the initial registration parameter.
9. the surgical navigational method based on the multi-modality images registration according to claim 8; It is characterized in that said nuclear magnetic resonance image in DRR image and the art is sampled according to the first sampling ratio obtained DRR image with first resolution and the step of the nuclear magnetic resonance image in the art specifically comprises:
The grey scale pixel value of the nuclear magnetic resonance image in said DRR image and the art is carried out normalization;
With the said first sampling ratio 256 gray levels are sampled.
10. the surgical navigational method based on the multi-modality images registration according to claim 8 is characterized in that, the sampling and the step of registration repeat 1 time, wherein, and according to reference picture and treat that the resolution of registering images is provided with the first sampling ratio and repeated sampling ratio.
11. the surgical navigational method based on the multi-modality images registration according to claim 10; It is characterized in that; Adopt gradient descent algorithm to the registration optimized Algorithm, step-size in search, the relaxation factor in the condition of convergence, maximum iteration time, the B-spline surface size of first sampling and repeated sampling is set.
12. the surgical navigational method based on the multi-modality images registration according to claim 8 is characterized in that, in the step that said calculating registration is estimated:
Adopt normalized mutual information value to estimate as registration.
13. the surgical navigational method based on the multi-modality images registration according to claim 8; It is characterized in that; Registration is estimated when reaching predetermined threshold value, and the nuclear magnetic resonance image in said DRR image and the art is carried out absolute registration, wherein definition B batten control mesh in absolute registration.
14. the surgical navigational method based on the multi-modality images registration according to claim 8 is characterized in that, saidly combines the undergo surgery step of navigation of the nuclear magnetic resonance image in the CT image and art before the said art to comprise according to said coordinate transformation parameter:
CT image before the art that similarity is maximum and the nuclear magnetic resonance image in the art carry out image co-registration according to the space matching relationship, and according to the Image Acquisition surgical navigational path after the said fusion.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110444772.0A CN102592137B (en) | 2011-12-27 | 2011-12-27 | Multi-modality image registration method and operation navigation method based on multi-modality image registration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110444772.0A CN102592137B (en) | 2011-12-27 | 2011-12-27 | Multi-modality image registration method and operation navigation method based on multi-modality image registration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102592137A true CN102592137A (en) | 2012-07-18 |
| CN102592137B CN102592137B (en) | 2014-07-02 |
Family
ID=46480747
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110444772.0A Active CN102592137B (en) | 2011-12-27 | 2011-12-27 | Multi-modality image registration method and operation navigation method based on multi-modality image registration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102592137B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020976A (en) * | 2012-12-31 | 2013-04-03 | 中国科学院合肥物质科学研究院 | Method and system for registering three-dimensional medical images on basis of weighted fuzzy mutual information |
| CN103345774A (en) * | 2013-07-17 | 2013-10-09 | 中国人民解放军第三军医大学 | Method for building three-dimensional multi-scale vectorization model |
| CN103479397A (en) * | 2013-10-15 | 2014-01-01 | 哈尔滨工业大学 | Method for extracting color ultrasound parametric blood flow signal based on relaxation algorithm |
| CN106204561A (en) * | 2016-07-04 | 2016-12-07 | 西安电子科技大学 | Prostate multi-modality images non-rigid registration method based on mixed model |
| CN107742099A (en) * | 2017-09-30 | 2018-02-27 | 四川云图睿视科技有限公司 | A kind of crowd density estimation based on full convolutional network, the method for demographics |
| CN109767461A (en) * | 2018-12-28 | 2019-05-17 | 上海联影智能医疗科技有限公司 | Medical image registration method, device, computer equipment and storage medium |
| CN109767458A (en) * | 2018-12-21 | 2019-05-17 | 西北大学 | A kind of sequential optimization method for registering of semi-automatic segmentation |
| CN110223220A (en) * | 2019-06-14 | 2019-09-10 | 北京百度网讯科技有限公司 | A kind of method and apparatus handling image |
| CN110415279A (en) * | 2019-06-25 | 2019-11-05 | 北京全域医疗技术集团有限公司 | Method for registering images, device and equipment |
| CN110473233A (en) * | 2019-07-26 | 2019-11-19 | 上海联影智能医疗科技有限公司 | Method for registering, computer equipment and storage medium |
| CN112348082A (en) * | 2020-11-06 | 2021-02-09 | 上海依智医疗技术有限公司 | Deep learning model construction method, image processing method and readable storage medium |
| CN112581606A (en) * | 2020-12-25 | 2021-03-30 | 江苏集萃复合材料装备研究所有限公司 | Spinal pedicle screw implantation operation navigation method |
| WO2023283795A1 (en) * | 2021-07-12 | 2023-01-19 | Shanghaitech University | Method for high-resolution image reconstruction |
| CN116071401A (en) * | 2023-01-28 | 2023-05-05 | 中日友好医院(中日友好临床医学研究所) | Virtual CT image generation method and device based on deep learning |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529188B (en) * | 2016-11-25 | 2019-04-19 | 苏州国科康成医疗科技有限公司 | Image processing method applied to surgical navigational |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1975781A (en) * | 2006-11-29 | 2007-06-06 | 南方医科大学 | Image elastic registrating method based on limited sampling global optimisation |
-
2011
- 2011-12-27 CN CN201110444772.0A patent/CN102592137B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1975781A (en) * | 2006-11-29 | 2007-06-06 | 南方医科大学 | Image elastic registrating method based on limited sampling global optimisation |
Non-Patent Citations (5)
| Title |
|---|
| 《黑龙江大学工程学报》 20110228 金英 《优化的CT图像配准方法的研究与比较》 93-95页 1-14 第2卷, 第1期 * |
| 庞皓文等: "《基于B样条的弹性配准在放疗CT图像间配准的应用》", 《中国医学物理学杂志》, vol. 28, no. 5, 30 September 2011 (2011-09-30), pages 2887 - 2889 * |
| 张红颖等: "《基于层次B样条的医学图像弹性配准方法》", 《天津大学学报》, 31 January 2007 (2007-01-31), pages 37 - 38 * |
| 杨劲，卓彦，陈霖: "《多分辨率的自动MRI图像配准》", 《四川联合大学学报（工程科学版）》, vol. 2, no. 3, 31 May 1998 (1998-05-31), pages 68 - 77 * |
| 金英: "《优化的CT图像配准方法的研究与比较》", 《黑龙江大学工程学报》, vol. 2, no. 1, 28 February 2011 (2011-02-28), pages 93 - 95 * |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020976A (en) * | 2012-12-31 | 2013-04-03 | 中国科学院合肥物质科学研究院 | Method and system for registering three-dimensional medical images on basis of weighted fuzzy mutual information |
| CN103345774A (en) * | 2013-07-17 | 2013-10-09 | 中国人民解放军第三军医大学 | Method for building three-dimensional multi-scale vectorization model |
| CN103345774B (en) * | 2013-07-17 | 2015-08-05 | 中国人民解放军第三军医大学 | A kind of modeling method of three-dimensional multi-scale vector quantization |
| CN103479397A (en) * | 2013-10-15 | 2014-01-01 | 哈尔滨工业大学 | Method for extracting color ultrasound parametric blood flow signal based on relaxation algorithm |
| CN103479397B (en) * | 2013-10-15 | 2015-01-07 | 哈尔滨工业大学 | Method for extracting color ultrasound parametric blood flow signal based on relaxation algorithm |
| CN106204561A (en) * | 2016-07-04 | 2016-12-07 | 西安电子科技大学 | Prostate multi-modality images non-rigid registration method based on mixed model |
| CN106204561B (en) * | 2016-07-04 | 2019-07-30 | 西安电子科技大学 | Prostate multi-modality images non-rigid registration method based on mixed model |
| CN107742099A (en) * | 2017-09-30 | 2018-02-27 | 四川云图睿视科技有限公司 | A kind of crowd density estimation based on full convolutional network, the method for demographics |
| CN109767458B (en) * | 2018-12-21 | 2023-01-20 | 西北大学 | Semi-automatic segmented sequential optimization registration method |
| CN109767458A (en) * | 2018-12-21 | 2019-05-17 | 西北大学 | A kind of sequential optimization method for registering of semi-automatic segmentation |
| CN109767461B (en) * | 2018-12-28 | 2021-10-22 | 上海联影智能医疗科技有限公司 | Medical image registration method and device, computer equipment and storage medium |
| CN109767461A (en) * | 2018-12-28 | 2019-05-17 | 上海联影智能医疗科技有限公司 | Medical image registration method, device, computer equipment and storage medium |
| CN110223220A (en) * | 2019-06-14 | 2019-09-10 | 北京百度网讯科技有限公司 | A kind of method and apparatus handling image |
| CN110415279A (en) * | 2019-06-25 | 2019-11-05 | 北京全域医疗技术集团有限公司 | Method for registering images, device and equipment |
| CN110473233A (en) * | 2019-07-26 | 2019-11-19 | 上海联影智能医疗科技有限公司 | Method for registering, computer equipment and storage medium |
| CN110473233B (en) * | 2019-07-26 | 2022-03-01 | 上海联影智能医疗科技有限公司 | Registration method, computer device, and storage medium |
| CN112348082A (en) * | 2020-11-06 | 2021-02-09 | 上海依智医疗技术有限公司 | Deep learning model construction method, image processing method and readable storage medium |
| CN112581606A (en) * | 2020-12-25 | 2021-03-30 | 江苏集萃复合材料装备研究所有限公司 | Spinal pedicle screw implantation operation navigation method |
| CN112581606B (en) * | 2020-12-25 | 2024-03-12 | 江苏集萃复合材料装备研究所有限公司 | Navigation method for vertebral pedicle screw implantation operation |
| WO2023283795A1 (en) * | 2021-07-12 | 2023-01-19 | Shanghaitech University | Method for high-resolution image reconstruction |
| CN116071401A (en) * | 2023-01-28 | 2023-05-05 | 中日友好医院(中日友好临床医学研究所) | Virtual CT image generation method and device based on deep learning |
| CN116071401B (en) * | 2023-01-28 | 2023-08-01 | 中日友好医院(中日友好临床医学研究所) | Virtual CT image generation method and device based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102592137B (en) | 2014-07-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102592137B (en) | Multi-modality image registration method and operation navigation method based on multi-modality image registration | |
| CN111161326B (en) | System and method for unsupervised deep learning of deformable image registration | |
| Rivaz et al. | Automatic deformable MR-ultrasound registration for image-guided neurosurgery | |
| Fischer et al. | A unified approach to fast image registration and a new curvature based registration technique | |
| Pluim et al. | Mutual-information-based registration of medical images: a survey | |
| JP5584441B2 (en) | Association method between volume data | |
| CN106934821B (en) | Conical beam CT and CT image registration method based on ICP algorithm and B spline | |
| JP4885138B2 (en) | Method and system for motion correction in a sequence of images | |
| JP2009536857A5 (en) | Deformable registration of images for image-guided radiation therapy | |
| JP2016517759A (en) | Method and system for mesh segmentation and mesh registration | |
| CN106952285B (en) | Lung motion estimation method based on prior statistical motion model and adaptive registration | |
| CN104867104A (en) | Method for obtaining anatomical structural atlas for target mouse based on XCT image non-rigid registration | |
| Vásquez Osorio et al. | Accurate CT/MR vessel‐guided nonrigid registration of largely deformed livers | |
| CN103295234A (en) | Medical image segmentation system and medical image segmentation method based on deformation surface models | |
| Ayatollahi et al. | A new hybrid particle swarm optimization for multimodal brain image registration | |
| AlZu'bi et al. | Transferable HMM probability matrices in multi‐orientation geometric medical volumes segmentation | |
| CN113506331A (en) | Method, apparatus, computer device and storage medium for registering tissue and organ | |
| Noorda et al. | Subject-specific four-dimensional liver motion modeling based on registration of dynamic MRI | |
| Zhou et al. | N-net: 3d fully convolution network-based vertebrae segmentation from ct spinal images | |
| Khmelinskii et al. | Atlas-based organ & bone approximation for ex-vivo μMRI mouse data: a pilot study | |
| Markel et al. | A 4D biomechanical lung phantom for joint segmentation/registration evaluation | |
| CN103985111A (en) | 4D-MRI super-resolution reconstruction method based on double-dictionary learning | |
| EP3234917B1 (en) | Method and system for calculating a displacement of an object of interest | |
| Huesman et al. | Deformable registration of multimodal data including rigid structures | |
| Ijaz et al. | Optimization strategies for ultrasound volume registration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20201030 Address after: No.6-10, 5th floor, building 1, No.1, Tidu street, Qingyang District, Chengdu, Sichuan 610000 Patentee after: Beijing fish claw Network Technology Co.,Ltd. Chengdu branch Address before: 1068 No. 518055 Guangdong city in Shenzhen Province, Nanshan District City Xili University School Avenue Patentee before: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY CHINESE ACADEMY OF SCIENCES Effective date of registration: 20201030 Address after: 226000 No. 188 Shijiang Road, South District of Shigang Town Industrial Park, Tongzhou District, Nantong City, Jiangsu Province Patentee after: NANTONG KANGSHENG MEDICAL EQUIPMENT Co.,Ltd. Address before: 610000 no.6-10, 5th floor, building 1, Tidu street, Qingyang District, Chengdu City, Sichuan Province 610000 Patentee before: Beijing fish claw Network Technology Co.,Ltd. Chengdu branch |