CN105701778B - A kind of method that metal artifacts are removed in the image from CT - Google Patents
A kind of method that metal artifacts are removed in the image from CT Download PDFInfo
- Publication number
- CN105701778B CN105701778B CN201610013374.6A CN201610013374A CN105701778B CN 105701778 B CN105701778 B CN 105701778B CN 201610013374 A CN201610013374 A CN 201610013374A CN 105701778 B CN105701778 B CN 105701778B
- Authority
- CN
- China
- Prior art keywords
- image
- projection
- original
- picture
- metal
- 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.)
- Active
Links
- 229910052751 metal Inorganic materials 0.000 title claims abstract description 111
- 239000002184 metal Substances 0.000 title claims abstract description 111
- 238000000034 method Methods 0.000 title claims abstract description 53
- 230000003044 adaptive effect Effects 0.000 claims abstract description 5
- 210000004872 soft tissue Anatomy 0.000 claims description 33
- 238000001883 metal evaporation Methods 0.000 claims description 26
- 210000001519 tissue Anatomy 0.000 claims description 24
- 230000011218 segmentation Effects 0.000 claims description 16
- 238000003709 image segmentation Methods 0.000 claims description 13
- 238000001914 filtration Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 abstract description 22
- 239000007943 implant Substances 0.000 description 13
- 230000000694 effects Effects 0.000 description 8
- 210000000988 bone and bone Anatomy 0.000 description 6
- 230000008030 elimination Effects 0.000 description 6
- 238000003379 elimination reaction Methods 0.000 description 6
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 5
- 239000010931 gold Substances 0.000 description 5
- 229910052737 gold Inorganic materials 0.000 description 5
- 238000002595 magnetic resonance imaging Methods 0.000 description 5
- 150000002739 metals Chemical class 0.000 description 5
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 4
- 238000005481 NMR spectroscopy Methods 0.000 description 4
- 238000012805 post-processing Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 238000007689 inspection Methods 0.000 description 3
- 239000010935 stainless steel Substances 0.000 description 3
- 229910001220 stainless steel Inorganic materials 0.000 description 3
- 229910001069 Ti alloy Inorganic materials 0.000 description 2
- RTAQQCXQSZGOHL-UHFFFAOYSA-N Titanium Chemical compound [Ti] RTAQQCXQSZGOHL-UHFFFAOYSA-N 0.000 description 2
- 229910045601 alloy Inorganic materials 0.000 description 2
- 239000000956 alloy Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 229910052742 iron Inorganic materials 0.000 description 2
- 238000002558 medical inspection Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000000638 solvent extraction Methods 0.000 description 2
- 239000010936 titanium Substances 0.000 description 2
- 229910052719 titanium Inorganic materials 0.000 description 2
- CWYNVVGOOAEACU-UHFFFAOYSA-N Fe2+ Chemical compound [Fe+2] CWYNVVGOOAEACU-UHFFFAOYSA-N 0.000 description 1
- BQCADISMDOOEFD-UHFFFAOYSA-N Silver Chemical compound [Ag] BQCADISMDOOEFD-UHFFFAOYSA-N 0.000 description 1
- 210000001015 abdomen Anatomy 0.000 description 1
- 210000003445 biliary tract Anatomy 0.000 description 1
- 238000005266 casting Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000002247 constant time method Methods 0.000 description 1
- 210000004351 coronary vessel Anatomy 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000003702 image correction Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 229910052758 niobium Inorganic materials 0.000 description 1
- 239000010955 niobium Substances 0.000 description 1
- GUCVJGMIXFAOAE-UHFFFAOYSA-N niobium atom Chemical compound [Nb] GUCVJGMIXFAOAE-UHFFFAOYSA-N 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000005311 nuclear magnetism Effects 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 210000002307 prostate Anatomy 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 229910052709 silver Inorganic materials 0.000 description 1
- 239000004332 silver Substances 0.000 description 1
- 210000003625 skull Anatomy 0.000 description 1
- GOLXNESZZPUPJE-UHFFFAOYSA-N spiromesifen Chemical compound CC1=CC(C)=CC(C)=C1C(C(O1)=O)=C(OC(=O)CC(C)(C)C)C11CCCC1 GOLXNESZZPUPJE-UHFFFAOYSA-N 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
- 238000007631 vascular surgery Methods 0.000 description 1
Classifications
-
- G06T5/77—
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
Abstract
The present invention relates to a kind of method that metal artifacts are removed in image from CT, pre-processed first by image adaptive filter, obtain eliminating the original reconstructed image of noise and part strip artifact；Then original reconstructed image is split by the method for cluster, obtains the region of different tissues, and establish model image, while the metallic region being partitioned into is obtained into its position in projection domain by orthographic projection；Next model image is obtained into its data for projection by orthographic projection, the projection numeric field data of the position model image according to the metallic region previously obtained in projection domain substitutes the projection numeric field data of original reconstructed image afterwards；The projection numeric field data after reparation is finally obtained into final correction chart picture by filtered back projection.The present invention accurately reduces true picture, can effectively remove metal artifacts, contribute to accurate judgement of the doctor to the state of an illness.
Description
Technical field
The present invention relates to the enhancing or recovery of CT images, and in particular to the side of metal artifacts is removed in a kind of image from CT
Method.
Background technology
Medically applied metal implant repairs human body the history for having more than 400 years, very early when doctor
Just repaired skull with proof gold plate and had a denture made, use silver, iron, ferrous alloy etc. successively thereafter to fix fracture joint, modern age
Cobalt-base alloys has been used also to be widely used as body implant, subsequent stainless steel again, with the development of rare metals industry,
The rare metal such as titanium and casting titanium, niobium is processed by clinical practice in human body, nowadays titanium alloy is since its intensity is high, corrosion resistance
Good, the features such as heat resistance is high and be reserved as the first choice of body implant, be widely used.
MRI is a kind of biological magnetic spin imaging technique, it is the characteristics of utilizing nuclear spin campaign, in externally-applied magnetic field
It is interior, signal is produced through radio-frequency pulse laser, is detected with detector and inputs computer, in screen after computer disposal is changed
Upper display image.
It is orthopaedics that magnetic resonance imaging (MRI), which checks, especially the effective tool of spinal surgery medical diagnosis on disease.
But have two following drawbacks when carrying out nuclear-magnetism inspection there are the human body of metal implant：
1) due to there is very powerful magnetic field in nuclear magnetic resonance machine and nuclear magnetic resonance check room, equipped with the heart
Dirty pacemaker person, and metal clip, metallic support person are left after vascular surgery, or other coronary arteries, oesophagus, prostate,
Biliary tract carry out metallic support operator, definitely forbid to make nuclear magnetic resonance check, otherwise, due to metal by the attraction of strong magnetic field and
It is mobile, it would be possible to produce serious consequence so that life danger.
2) for the patient with metal implant, when inspection, may occur in which a large amount of puppets in MRI image metal implant area
Shadow, so that influencing diagnostic imaging performance.Orthopaedics implant material is used as using iron, stainless steel etc., artifact is fairly obvious, serious dry
The quality of image is disturbed.
Therefore the patient containing metal does not do nuclear magnetic resonance check, especially internal gold without exception in some hospital formulary bodies
Belong to the situation that implant is stainless steel rather than titanium alloy.Just because of MRI clinical practice as above the drawbacks of, CT is as another
Kind medical inspection means just have obvious advantage.
CT has as conventional medical inspection means in the case where patient's body has metal implant than MRI preferably should
With, but since the presence of metal implant still can have a certain impact the reconstructed image that it is scanned.Typical situation is just
It is that the presence of metal object make it that the image after rebuilding produces substantial amounts of black in metallic perimeter and bright radial-like streak is pseudo-
Shadow, judgement of the doctor to inspection result can be seriously affected when metal implant is bigger.
The main following points of metal artifacts Producing reason：When being implanted the larger metal of density in human body, due to
Its attenuation coefficient is far longer than the attenuation coefficient of its hetero-organization human body Nei, will largely weaken when ray passes through metal, produces
Beam hardening phenomenon causes the first derivative of data for projection to show weak flatness at a certain section, produces the jump of data for projection,
After filtered processing, this weak flatness can be further magnified, so as to finally be formed in reconstruction image light and dark
Strip artifact.Simultaneously because X-ray hardens problem, non-linear partial volume effect is produced, and scattering phenomenon can be aggravated, this
A little distortions that can all cause reconstructed image, especially form substantial amounts of interference in metallic perimeter.
In order to recover the reconstructed image polluted by metal artifacts, substantial amounts of metal artifacts elimination algorithm is suggested.
Method main at present is segmented into following three types：1. carrying out the elimination of metal artifacts with iterative reconstruction algorithm, 2. in projection domain
Carry out the elimination of metal artifacts.3. carry out the elimination of metal artifacts in image area.
Iterative reconstruction algorithm be theoretically it is extraordinary go metal artifacts method, the algorithm can effectively remove metal artifacts and
Suppress noise, and the structure of metal object can be presented well, but its operand is very big, and speed is very slow, it is difficult to practical.
Projection domain, which removes metal artifacts, to be a kind of method of pre-treatment and has certain advantage, because after image area removes metal artifacts
Processing method is needed by backprojection algorithms such as FBP, this can increase influence of the metal artifacts to truthful data, and include metal
For the reconstructed image of artifact since it has largely been polluted, larger error will necessarily be brought by carrying out processing on this basis.But
It is that current mainstream algorithm is most of or the post-processing algorithm based on reconstructed image, reason mainly have following two aspects：1.
Original raw data are not required in post-processing approach based on reconstructed image, it is only necessary to which the reconstructed image of DICOM format can complete gold
Belong to the elimination of artifact, data acquisition is relatively easy.2. the pre-treating method based on projection domain there are many technological difficulties, project first
Data volume is very big, is difficult secondly to have accurate method to split metal part from other human body parts in projection domain,
Therefore it is difficult the data for projection that Exact recovery is contaminated with metals, has its advantage so while carrying out metal artifacts from projection domain and eliminating
But its application and the post-processing algorithm for being not so good as image area.
The image area post-processing approach of metal artifacts is broadly divided into following three types：1. 2. filtration combined weighted correction of interpolation correction method
3. model correction method of method.Interpolation correction method is most known method, its main thought is the data for projection interpolation with metallic perimeter
Go out the data for projection being contaminated with metals, followed by filtered back projection obtains reconstructed image, the metal portion being finally partitioned into again
Partial image, which is added to, obtains final correction chart picture in reconstructed image.This method is simple and practical, can have to metal artifacts certain
Elimination acts on, but due to it simply recovers the data for projection being contaminated with metals simply by interpolation method, does not consider
The details of metal part, and even if selection high-order interpolation function also is difficult to ensure flatness of the data for projection in metal edge,
The metal part image superposition for finally needing to be partitioned into additionally, due to its algorithm is into final reconstructed image so that in metal area
The edge in domain can not have well excessively, therefore interpolation correction method is unsatisfactory to the treatment effect of metal artifacts.Filtering adds
Power method is that a variety of known images gone after metal artifacts method (such as nmar) correction are carried out high pass and low-pass filtering treatment,
Then fusion is weighted, so as to obtain final correction chart picture, which can obtain certain effect, but compares and rely on it
With basis remove artifact algorithm, actual effect is unsatisfactory.Model correction method is bearing calibration popular in recent years, its
Basic thought is preferably to reduce true picture by establishing model image, and the data for projection of model image is substituted original throwing
The part being contaminated with metals in shadow data, finally obtains the reconstructed image of no metallic pollution.This method computation complexity is low, and
And retain the details of metal while removal metal artifacts that can be relatively good, it is especially excessively natural in metal edge.Should
The difficult point of method is how to establish accurate model image to reduce real image, this is also the root of different model correcting algorithms
Where this difference.
Model correction method has than interpolation correction method and the more preferable accuracy of filtration combined weighted method and practicality, so the present invention
It is metal artifacts to be removed based on model correction method.Model correction method is how accurately to establish illustraton of model at all
Picture, this difference place for being both different model correction methods are even more the difficult point place of such correction method.
More accurate model bearing calibration is proposed in order to effectively remove the metal artifacts present invention, this method is in model
During the foundation of image, the minutia of different tissues and the characteristic of metal object have been taken into full account, not only fully application
The advantage of model correction method, has more carried out improve and perfect, so as to obtain more preferable gold according to actual clinical data to it
Belong to artifact removal effect.
The content of the invention
The object of the present invention is to provide a kind of method that metal artifacts are removed in image from CT, pass through image adaptive first
Filtering is pre-processed, and obtains eliminating the original reconstructed image of noise and part strip artifact；Then the method for cluster is passed through
Original reconstructed image is split, obtains the region of different tissues, and establishes model image, while the metal area that will be partitioned into
Domain obtains its position in projection domain by orthographic projection；Next model image is obtained into its data for projection by orthographic projection,
The projection numeric field data of the position model image according to the metallic region previously obtained in projection domain substitutes original reconstructed afterwards
The projection numeric field data of image；The projection numeric field data after reparation is finally obtained into final correction chart picture by filtered back projection.
The purpose of the present invention is what is realized by following technical proposals：A kind of method that metal artifacts are removed in the image from CT,
The step of removal metal artifacts, includes：
A, image preprocessing：The noise and part strip artifact in CT images are removed using self-adaptive routing, obtains original
Beginning reconstructed image；
B, image is split：Split the original reconstructed image using cluster segmentation algorithm, obtain including different tissues region
Original dendrogram picture；The tissue regions include metallic region, bony areas, soft tissue area, air section；
C, weight map picture is established：Calculate diverse location and metal in the original dendrogram picture in the original dendrogram picture
The distance between region, the weighted value of the position pixel is calculated according to the distance, obtains weight map picture；The weight map picture is not by
Weighted value with position pixel is formed；
D, image segmentation is corrected：The weighted value of pixel in the original dendrogram picture is more than in advance according to the weight map picture
The air section for the threshold value being first set in the range of [0,1] is modified to soft tissue area, obtains correcting dendrogram picture；
E, model image is established：Ask for the soft tissue area pair with the amendment dendrogram picture in the original reconstructed image
Answer the average value of all pixels in region, and all pictures with the average value to the corresponding region in the original reconstructed image
Element carries out assignment, obtains model image；
F, the projection domain position of metallic region is determined：Metallic region in the original dendrogram picture is subjected to orthographic projection,
Metal shadowing's area image is obtained, the location of pixels being not zero in metal shadowing's area image is recorded, obtains the metallic region
Projection domain position in metal shadowing's area image；
G, the projection area image of original reconstructed image is corrected based on model image：Respectively by the original reconstructed image and institute
State model image and carry out orthographic projection, obtain the projection area image of the original reconstructed image and the projection domain figure of the model image
Picture；According to projection domain position of the metallic region in metal shadowing's area image, with the projection domain of the model image
Corresponded in image in the projection area image of the projection numeric field data replacement original reconstructed image of the projection domain position and correspond to the throwing
The projection numeric field data of shadow domain position, obtains the amendment projection area image of the original reconstructed image；
H, the amendment projection area image of the original reconstructed image obtains removing repairing for metal artifacts by filtered back projection
Positive reconstructed image.
The purpose of the present invention can also be realized by following technical proposals：The side of metal artifacts is removed in a kind of image from CT
The step of method, the removal metal artifacts, includes：
A, image preprocessing：The noise and part strip artifact in CT images are removed using self-adaptive routing, obtains original
Beginning reconstructed image；
B, image is split：Split the original reconstructed image using cluster segmentation algorithm, obtain including different tissues region
Original dendrogram picture；The tissue regions include metallic region, bony areas, soft tissue area, air section；
C, the projection domain position of metallic region is determined：Metallic region in the original dendrogram picture is subjected to orthographic projection,
Metal shadowing's area image is obtained, the location of pixels being not zero in metal shadowing's area image is recorded, obtains the metallic region
Projection domain position in metal shadowing's area image；
D, weight map picture is established：Calculate diverse location and metal in the original dendrogram picture in the original dendrogram picture
The distance between region, the weighted value of the position pixel is calculated according to the distance, obtains weight map picture；The weight map picture is not by
Weighted value with position pixel is formed；
E, image segmentation is corrected：The weighted value of pixel in the original dendrogram picture is more than in advance according to the weight map picture
The air section for the threshold value being first set in the range of [0,1] is modified to soft tissue area, obtains correcting dendrogram picture；
F, model image is established：Ask for the soft tissue area pair with the amendment dendrogram picture in the original reconstructed image
Answer the average value of all pixels in region, and all pictures with the average value to the corresponding region in the original reconstructed image
Element carries out assignment, obtains model image；
G, the projection area image of original reconstructed image is corrected based on model image：Respectively by the original reconstructed image and institute
State model image and carry out orthographic projection, obtain the projection area image of the original reconstructed image and the projection domain figure of the model image
Picture；According to projection domain position of the metallic region in metal shadowing's area image, with the projection domain of the model image
Corresponded in image in the projection area image of the projection numeric field data replacement original reconstructed image of the projection domain position and correspond to the throwing
The projection numeric field data of shadow domain position, obtains the amendment projection area image of the original reconstructed image；
H, the amendment projection area image of the original reconstructed image obtains removing repairing for metal artifacts by filtered back projection
Positive reconstructed image.
Further, the self-adaptive routing is median filtering method.
Further, the cluster segmentation algorithm is K-means cluster segmentation algorithms.
Further, the tissue regions further include normal tissue.
The present invention has the following advantages that compared with prior art：
1st, current model correction method generally be simple structure model image, without considering metal implant feature and
Position in human body, and simply simple threshold application method splits different tissue regions, the model image mistake of foundation
In simple inaccurate.The present invention has taken into full account different parts institute issuable influence of the metal implant in human body, and
Adaptive clustering algorithm is applied for different body parts, so as to establish more accurate model image, is finally obtained
More accurate correction chart picture.
2nd, the reconstructed image that finally obtains of the present invention is very effective on the whole eliminate metal artifacts caused by
Bright band and blanking bar, and accurately reduce true picture also there is no obvious hole around metal, illustrate improved
Model image correction method can effectively remove metal artifacts, contribute to accurate judgement of the doctor to the state of an illness.
The present invention is described in detail below in conjunction with the drawings and specific embodiments.
Brief description of the drawings
Fig. 1 is the flow chart of present pre-ferred embodiments；
Fig. 2 is CT images；
Fig. 3 is original reconstructed image；
Fig. 4 is original dendrogram picture；
Fig. 5 is weight map picture；
Fig. 6 is to correct dendrogram picture；
Fig. 7 is model image；
Fig. 8 is metal shadowing's area image；
Fig. 9 is the amendment reconstructed image for removing metal artifacts.
Embodiment
Referring to Fig. 1, a kind of method of removal metal artifacts in image from CT, described the step of removing metal artifacts, includes：
A, image preprocessing：The noise and part strip artifact in CT images are removed using self-adaptive routing, obtains original
Beginning reconstructed image；The present invention eliminates partial noise using medium filtering, however, it would be possible to use more complicated filtering method comes
Preferably eliminate noise.
Fig. 2 is CT original images, and CT original images are exactly the CT reconstructed images not handled by eliminating metal artifacts 1,
The image of traditional sense is may be substantially identical to, only form may some difference (dicom etc.).Fig. 3 is adaptively to filter
Image (due to original image noise very little so after pretreatment image and original image difference and unobvious) after ripple.
B, image is split：Split the original reconstructed image using cluster segmentation algorithm, obtain including different tissues region
Original dendrogram picture；The tissue regions include metallic region, bony areas, soft tissue area, air section；
Referring to Fig. 4, image segmentation stage：Sweep object can be substantially divided into according to the structure of human body by metal (A) 2,
The different tissue regions such as bone (B) 3, soft tissue (C) 4, air (D) 5, also may be used with the different classifications split of scanned position
With change, for example during head of the scanning comprising metal tooth, sweep object can be divided into metal (A), bone (B), soft group
Knit four classes such as (C) and air (D)；But when scanning belly when position, then sweep object can be divided into metal (A), bone
Bone (B), soft tissue (C1), five classes such as normal tissue (C2) and air (D), can so obtain finer segmentation, so as to establish
Finer model image.Different classification sums primarily to can be finer accurate in the segmentation stage, but it is final I
Sweep object is all divided into four classes --- soft tissue and normal tissue (are merged into one by-metal, bone, soft tissue and air
Class).Partitioning algorithm does not limit in principle, it is only necessary to obtain accurate segmentation result can, the present invention selection K-means gather
Class partitioning algorithm.
It can be seen that sweep object has been divided into four classes in Fig. 4, different colors is corresponded to respectively.(pay attention to metallic perimeter
Air section, actually should not be existing)
I (i, j) is original segmentation figure picture, I'(i, j) for the segmentation figure picture after merging, when classification sum is four,
I (i, j) and I'(i, j) it is identical, when classification sum is five, I'(i, j) it is calculated by above formula.
C, weight map picture is established：Calculate diverse location and metal in the original dendrogram picture in the original dendrogram picture
The distance between region, the weighted value of the position pixel is calculated according to the distance, obtains weight map picture；The weight map picture is not by
Weighted value with position pixel is formed；
The common influence of metal artifacts is all near metallic region, and closer metallic region artifact is more serious, so having
The judgement current pixel of effect is extremely important by metal effect, this allow we focus more on processing those by metal
The big region of artifact effects, this aspect cause the more targeted of processing, on the other hand can also reduce to the greatest extent due to correction
Caused error.The present invention is by calculating the distance of different images position and metallic region come for diverse location in reconstructed image
Pixel calculate weight so as to judge its size influenced by metal artifacts.The standard selection Euclidean distance of distance is (in principle
It is not limited to Euclidean distance), obtain being inversely proportional with distance apart from the calculating of rear weight, that is, after normalization inside metallic region
Weight is 1, smaller further away from metallic region weight, until 0.
Fig. 5 can be seen that the bigger position weight of brightness is higher, that is, nearer from metallic region.
Wherein, W (i, j) is weight map picture, and D (i, j) is that current location and the Euclidean distance of metallic region all the points are minimum
Value, maxDistance is set in advance apart from maximum.
D, image segmentation is corrected：The weighted value of pixel in the original dendrogram picture is more than in advance according to the weight map picture
The air section for the threshold value being first set in the range of [0,1] is modified to soft tissue area, obtains correcting dendrogram picture；
Referring to Fig. 6, image segmentation amendment stage：Since metal is all to be implanted to inside of human body, so around metal 2 usually
Should be soft tissue 4, and metal artifacts can produce blanking bar in metallic perimeter, these blanking bars probably split stage quilt in image
Mistake is assigned in air section 5, and this is not inconsistent with general knowledge, so we should assign it to soft tissue area again
In domain 4.According to the weight map picture obtained before, the image that can easily know which position be in metallic perimeter, so
Those very big regions for belonging to air of weight can be modified to soft tissue area by us, so as to obtain more accurate image point
Cut result.
Wherein Ic(i, j) is revised dendrogram picture, and threshold is the threshold value being set in advance in the range of [0,1].
Specifically, it is exactly that image can be divided into four parts (air, soft tissue, bone after original reconstructed image segmentation
Bone and metal), the part for being divided into air section, if the also very big (power of the weighted value of correspondence position in weight map picture
Weight values are in 0~1 scope, such as more than 0.9), then this partial air region is just divided into soft tissue area again, that is,
Original dendrogram picture is modified.
Fig. 6 can be seen that revised segmentation figure picture eliminates the hole near metallic region, obtain more accurate point
Cut image.
E, model image is established：Ask for the soft tissue area pair with the amendment dendrogram picture in the original reconstructed image
Answer the average value of all pixels in region, and all pictures with the average value to the corresponding region in the original reconstructed image
Element carries out assignment, obtains model image；
Model image establishment stage：Sweep object is divided into different parts after image segmentation, usually, gold
Belong to the CT values of artifact closer to soft tissue area, so can be divided among soft tissue area, it is more real in order to reduce
Scan image, when establishing model image, we carry out assignment for regional image again.Due to metal artifacts major part quilt
Soft tissue area is divided into, so the image of soft tissue area is averaged, and is reassigned to all pixels.Other areas
The numerical value in domain is constant.Have benefited from, We conducted the amendment of image segmentation, more accurate image segmentation result having been obtained, so energy
It is enough to establish more accurate model image.
Wherein IM(i, j) is model image, CTavgCFor the average of soft tissue area's CT values.
Fig. 7 can be seen that metal artifacts have been included in soft tissue area by the model image of foundation.
F, the projection domain position of metallic region is determined：Metallic region in the original dendrogram picture is subjected to orthographic projection,
Metal shadowing's area image is obtained, the location of pixels being not zero in metal shadowing's area image is recorded, obtains the metallic region
Projection domain position in metal shadowing's area image；
The metal shadowing domain location determination stage：The image segmentation stage is being carried out, we can obtain single metallic region
Image, can further obtain it by orthographic projection and project numeric field data, by judging which data is not zero, may finally obtain
Specific location (channel position each visual angle under) of the metallic region in numeric field data is projected.
Fig. 8 can be seen that the scope of metallic region in projection domain.
G, the projection area image of original reconstructed image is corrected based on model image：Respectively by the original reconstructed image and institute
State model image and carry out orthographic projection, obtain the projection area image of the original reconstructed image and the projection domain figure of the model image
Picture；According to projection domain position of the metallic region in metal shadowing's area image, with the projection domain of the model image
Corresponded in image in the projection area image of the projection numeric field data replacement original reconstructed image of the projection domain position and correspond to the throwing
The projection numeric field data of shadow domain position, obtains the amendment projection area image of the original reconstructed image；
Projection domain based on model image corrects the stage：After model image is obtained, respectively by original image and illustraton of model
As carrying out orthographic projection, corresponding projection numeric field data is obtained, according to position of the predetermined metallic region in projection domain, uses mould
The data for projection of type image replaces Raw projection data, so as to obtain revised data for projection.In model image data for projection
, can be with order to enable more smooth so as to the new artifact that do not induce one in metal edge during replacing original image data for projection
Metal boundary data for projection is handled using mean filter.
H, the amendment projection area image of the original reconstructed image obtains removing repairing for metal artifacts by filtered back projection
Positive reconstructed image.
After obtaining modified data for projection, final amendment reconstructed image is obtained according to filter back-projection algorithm.
Wherein, PC(k) it is the corrected projection data of some visual angle lower channel, PO(k) it is the original throwing of corresponding visual angle lower channel
Shadow data, PM(k) it is the model projection data of corresponding visual angle lower channel, m, n represent metal on the scope side of respective channel data
Boundary.
Fig. 9 can be seen that final correction chart as it is very effective on the whole eliminate metal artifacts caused by it is bright
Band and blanking bar, and accurately reduce true picture also there is no obvious hole around metal, illustrate improved mould
Type image calibration, which is executed, can effectively remove metal artifacts, contribute to accurate judgement of the doctor for the state of an illness.
In another embodiment, the step of removal metal artifacts include：
A, image preprocessing：The noise and part strip artifact in CT images are removed using self-adaptive routing, obtains original
Beginning reconstructed image；
B, image is split：Split the original reconstructed image using cluster segmentation algorithm, obtain including different tissues region
Original dendrogram picture；The tissue regions include metallic region, bony areas, soft tissue area, air section；
C, the projection domain position of metallic region is determined：Metallic region in the original dendrogram picture is subjected to orthographic projection,
Metal shadowing's area image is obtained, the location of pixels being not zero in metal shadowing's area image is recorded, obtains the metallic region
Projection domain position in metal shadowing's area image；
D, weight map picture is established：Calculate diverse location and metal in the original dendrogram picture in the original dendrogram picture
The distance between region, the weighted value of the position pixel is calculated according to the distance, obtains weight map picture；The weight map picture is not by
Weighted value with position pixel is formed；
E, image segmentation is corrected：The weighted value of pixel in the original dendrogram picture is more than in advance according to the weight map picture
The air section for the threshold value being first set in the range of [0,1] is modified to soft tissue area, obtains correcting dendrogram picture；
F, model image is established：Ask for the soft tissue area pair with the amendment dendrogram picture in the original reconstructed image
Answer the average value of all pixels in region, and all pictures with the average value to the corresponding region in the original reconstructed image
Element carries out assignment, obtains model image；
G, the projection area image of original reconstructed image is corrected based on model image：Respectively by the original reconstructed image and institute
State model image and carry out orthographic projection, obtain the projection area image of the original reconstructed image and the projection domain figure of the model image
Picture；According to projection domain position of the metallic region in metal shadowing's area image, with the projection domain of the model image
Corresponded in image in the projection area image of the projection numeric field data replacement original reconstructed image of the projection domain position and correspond to the throwing
The projection numeric field data of shadow domain position, obtains the amendment projection area image of the original reconstructed image；
H, the amendment projection area image of the original reconstructed image obtains removing repairing for metal artifacts by filtered back projection
Positive reconstructed image.
The content of the present embodiment is merely preferred embodiments of the present invention, but protection scope of the present invention is not limited to
In this, any one skilled in the art in the technical scope of present disclosure, the change that can readily occur in or
Replace, should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with the guarantor of claims
Protect subject to scope.
Claims (8)
1. the method for metal artifacts is removed in a kind of image from CT, it is characterised in that：The step of removal metal artifacts, includes：
A, image preprocessing：The noise and part strip artifact in CT images are removed using self-adaptive routing, is obtained original heavy
Composition picture；
B, image is split：Split the original reconstructed image using cluster segmentation algorithm, obtain including the original in different tissues region
Beginning dendrogram picture；The tissue regions include metallic region, bony areas, soft tissue area, air section；
C, weight map picture is established：Calculate diverse location and metallic region in the original dendrogram picture in the original dendrogram picture
The distance between, the weighted value of the position pixel is calculated according to the distance, obtains weight map picture；The weight map picture is by different positions
The weighted value for putting pixel is formed；
D, image segmentation is corrected：The weighted value of pixel in the original dendrogram picture is more than according to the weight map picture and is set in advance
The air section for the threshold value being scheduled in the range of [0,1] is modified to soft tissue area, obtains correcting dendrogram picture；
E, model image is established：Ask for area corresponding with the soft tissue area of the amendment dendrogram picture in the original reconstructed image
The average value of all pixels in domain, and with the average value to all pixels of the corresponding region in the original reconstructed image into
Row assignment, obtains model image；
F, the projection domain position of metallic region is determined：Metallic region in the original dendrogram picture is subjected to orthographic projection, is obtained
Metal shadowing's area image, records the location of pixels being not zero in metal shadowing's area image, obtains the metallic region in institute
State the projection domain position in metal shadowing's area image；
G, the projection area image of original reconstructed image is corrected based on model image：Respectively by the original reconstructed image and the mould
Type image carries out orthographic projection, obtains the projection area image of the original reconstructed image and the projection area image of the model image；
According to projection domain position of the metallic region in metal shadowing's area image, with the projection area image of the model image
The projection numeric field data of the middle correspondence projection domain position replaces in the projection area image of the original reconstructed image and corresponds to the projection domain
The projection numeric field data of position, obtains the amendment projection area image of the original reconstructed image；
H, the amendment projection area image of the original reconstructed image obtains removing the amendment weight of metal artifacts by filtered back projection
Composition picture.
2. the method for metal artifacts is removed in the image according to claim 1 from CT, it is characterised in that：Described is adaptive
Filter method is median filtering method.
3. the method for metal artifacts is removed in the image according to claim 1 from CT, it is characterised in that：The cluster point
It is K-means cluster segmentation algorithms to cut algorithm.
4. the method for metal artifacts is removed in the image according to claim 1 from CT, it is characterised in that：The tissue regions
Further include normal tissue.
5. the method for metal artifacts is removed in a kind of image from CT, it is characterised in that：The step of removal metal artifacts, includes：
A, image preprocessing：The noise and part strip artifact in CT images are removed using self-adaptive routing, is obtained original heavy
Composition picture；
B, image is split：Split the original reconstructed image using cluster segmentation algorithm, obtain including the original in different tissues region
Beginning dendrogram picture；The tissue regions include metallic region, bony areas, soft tissue area, air section；
C, the projection domain position of metallic region is determined：Metallic region in the original dendrogram picture is subjected to orthographic projection, is obtained
Metal shadowing's area image, records the location of pixels being not zero in metal shadowing's area image, obtains the metallic region in institute
State the projection domain position in metal shadowing's area image；
D, weight map picture is established：Calculate diverse location and metallic region in the original dendrogram picture in the original dendrogram picture
The distance between, the weighted value of the position pixel is calculated according to the distance, obtains weight map picture；The weight map picture is by different positions
The weighted value for putting pixel is formed；
E, image segmentation is corrected：The weighted value of pixel in the original dendrogram picture is more than according to the weight map picture and is set in advance
The air section for the threshold value being scheduled in the range of [0,1] is modified to soft tissue area, obtains correcting dendrogram picture；
F, model image is established：Ask for area corresponding with the soft tissue area of the amendment dendrogram picture in the original reconstructed image
The average value of all pixels in domain, and with the average value to all pixels of the corresponding region in the original reconstructed image into
Row assignment, obtains model image；
G, the projection area image of original reconstructed image is corrected based on model image：Respectively by the original reconstructed image and the mould
Type image carries out orthographic projection, obtains the projection area image of the original reconstructed image and the projection area image of the model image；
According to projection domain position of the metallic region in metal shadowing's area image, with the projection area image of the model image
The projection numeric field data of the middle correspondence projection domain position replaces in the projection area image of the original reconstructed image and corresponds to the projection domain
The projection numeric field data of position, obtains the amendment projection area image of the original reconstructed image；
H, the amendment projection area image of the original reconstructed image obtains removing the amendment weight of metal artifacts by filtered back projection
Composition picture.
6. the method for metal artifacts is removed in the image according to claim 5 from CT, it is characterised in that：Described is adaptive
Filter method is median filtering method.
7. the method for metal artifacts is removed in the image according to claim 5 from CT, it is characterised in that：The cluster point
It is K-means cluster segmentation algorithms to cut algorithm.
8. the method for metal artifacts is removed in the image according to claim 5 from CT, it is characterised in that：The tissue regions
Further include normal tissue.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610013374.6A CN105701778B (en) | 2016-01-11 | 2016-01-11 | A kind of method that metal artifacts are removed in the image from CT |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610013374.6A CN105701778B (en) | 2016-01-11 | 2016-01-11 | A kind of method that metal artifacts are removed in the image from CT |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105701778A CN105701778A (en) | 2016-06-22 |
| CN105701778B true CN105701778B (en) | 2018-04-20 |
Family
ID=56227036
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610013374.6A Active CN105701778B (en) | 2016-01-11 | 2016-01-11 | A kind of method that metal artifacts are removed in the image from CT |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105701778B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106447631B (en) * | 2016-09-19 | 2019-04-02 | 深圳市贝斯达医疗股份有限公司 | The processing method and processing device of image |
| CN106485287A (en) * | 2016-10-25 | 2017-03-08 | 帝麦克斯(苏州)医疗科技有限公司 | The acquisition methods of scanning area, the acquisition methods of digital slices and device |
| CN107316277B (en) * | 2017-04-24 | 2020-12-18 | 苏州动影信息科技有限公司 | Method for correcting motion artifact in medical cone beam CT image |
| CN107978003B (en) * | 2017-10-30 | 2021-03-30 | 东软医疗系统股份有限公司 | CT image metal artifact processing method and device |
| CN108961176B (en) * | 2018-06-14 | 2021-08-03 | 中国科学院半导体研究所 | Self-adaptive bilateral reference restoration method for range-gated three-dimensional imaging |
| CN109472754A (en) * | 2018-11-06 | 2019-03-15 | 电子科技大学 | CT image metal artifact removing method based on image repair |
| CN109859140A (en) * | 2019-02-15 | 2019-06-07 | 数坤（北京）网络科技有限公司 | A kind of restorative procedure and equipment for medical image |
| CN110097517B (en) * | 2019-04-28 | 2022-12-27 | 东软医疗系统股份有限公司 | Method and device for removing image artifacts |
| CN111402150B (en) * | 2020-03-09 | 2023-07-28 | 北京灵医灵科技有限公司 | CT image metal artifact removal method and device |
| CN111815521B (en) * | 2020-05-27 | 2024-02-13 | 南京国科精准医学科技有限公司 | Cone beam CT metal artifact correction algorithm based on priori image |
| CN111803107B (en) * | 2020-08-06 | 2023-09-12 | 上海联影医疗科技股份有限公司 | Metal detection method, device, equipment and storage medium |
| CN113256542B (en) * | 2021-06-22 | 2021-10-01 | 明峰医疗系统股份有限公司 | Noise suppression method, system and medium for CT scanner |
| CN113520441B (en) * | 2021-08-03 | 2022-11-29 | 浙江大学 | Tissue imaging method and system for eliminating CT high-impedance artifact interference |
| CN113902824B (en) * | 2021-10-12 | 2022-10-28 | 哈尔滨工业大学 | Guide wire artifact removing method for intravascular optical coherence tomography |
| CN114983452A (en) * | 2022-05-11 | 2022-09-02 | 赛诺威盛科技(北京)股份有限公司 | CT spiral reconstruction image artifact removing method and device |
| CN116309923A (en) * | 2023-05-24 | 2023-06-23 | 吉林大学 | CT metal artifact eliminating method and system based on graph neural network |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060285737A1 (en) * | 2005-06-17 | 2006-12-21 | Hamill James J | Image-based artifact reduction in PET/CT imaging |
| CN102376097A (en) * | 2010-08-25 | 2012-03-14 | 东软飞利浦医疗设备系统有限责任公司 | Computed tomography (CT) iteration reconstruction method for asymmetrical detector |
| CN102567958B (en) * | 2010-12-31 | 2014-06-25 | 飞利浦(中国)投资有限公司 | Image postprocessing method for removing metal artifact from computed tomography (CT) image |
| CN103020928B (en) * | 2012-11-21 | 2016-06-29 | 深圳先进技术研究院 | The metal artifacts reduction method of cone-beam CT system |
| CN104504655A (en) * | 2014-12-04 | 2015-04-08 | 沈阳东软医疗系统有限公司 | CT (computer tomography) metal artifact processing method and device |
-
2016
- 2016-01-11 CN CN201610013374.6A patent/CN105701778B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105701778A (en) | 2016-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105701778B (en) | A kind of method that metal artifacts are removed in the image from CT | |
| US5832134A (en) | Data visualization enhancement through removal of dominating structures | |
| JP4874810B2 (en) | X-ray CT system | |
| CN105225208B (en) | A kind of computer tomography metal artifacts reduction method and device | |
| US20090074278A1 (en) | Method and apparatus for metal artifact reduction in computed tomography | |
| US7330573B2 (en) | Visualization of medical images | |
| Meilinger et al. | Metal artifact reduction in cone beam computed tomography using forward projected reconstruction information | |
| CN102024251B (en) | System and method for multi-image based virtual non-contrast image enhancement for dual source CT | |
| CN104899876B (en) | A kind of eye fundus image blood vessel segmentation method based on adaptive Gauss difference | |
| JP2007014759A (en) | Method for reducing image-based artifact in pet/ct imaging | |
| CN106960429B (en) | A kind of CT image metal artifact bearing calibration and device | |
| CN112017131B (en) | CT image metal artifact removing method and device and computer readable storage medium | |
| EP2869261B1 (en) | Method for processing image data representing a three-dimensional volume | |
| DE102012206714A1 (en) | Method, arithmetic unit, CT system and C-arm system for the reduction of metal artifacts in CT image datasets | |
| Kesuma et al. | Improved Chest X-Ray Image Quality Using Median and Gaussian Filter Methods | |
| CN112562058B (en) | Method for quickly establishing intracranial vascular simulation three-dimensional model based on transfer learning | |
| Park et al. | MRI-based computed tomography metal artifact correction method for improving proton range calculation accuracy | |
| CN113034522A (en) | CT image segmentation method based on artificial neural network | |
| Kohan et al. | The impact of orthopedic metal artifact reduction software on interreader variability when delineating areas of interest in the head and neck | |
| CN102124471A (en) | Methods for enhancing vascular patterns in cervical imagery | |
| JPH08287230A (en) | Computer-aided image diagnostic device | |
| CN114387191B (en) | Endoscope image enhancement method and endoscope device | |
| Sayin et al. | Optimization of CBCT data with image processing methods and production with fused deposition modeling 3D printing | |
| Frosio et al. | Real time enhancement of cephalometric radiographies | |
| Schwemmer et al. | High-Density Object Removal from Projection Images using Low-Frequency-Based Object Masking. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Method of removing metal artifact from CT image Effective date of registration: 20180813 Granted publication date: 20180420 Pledgee: Beijing Zhongguancun bank Limited by Share Ltd Pledgor: Sainuo via science and Technology (Beijing) Co. Ltd. Registration number: 2018110000019 |
|
| 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: 20200323 Granted publication date: 20180420 Pledgee: Beijing Zhongguancun bank Limited by Share Ltd Pledgor: SAINUO WEISHENG TECHNOLOGY (BEIJING) Co.,Ltd. Registration number: 2018110000019 |
|
| PC01 | Cancellation of the registration of the contract for pledge of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Method of removing metal artifact from CT image Effective date of registration: 20200323 Granted publication date: 20180420 Pledgee: Beijing Zhongguancun bank Limited by Share Ltd Pledgor: SAINUO WEISHENG TECHNOLOGY (BEIJING) Co.,Ltd. Registration number: Y2020990000200 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| CP03 | Change of name, title or address |
Address after: 100176 1st floor, building 8, 11 Kangding street, Beijing Economic and Technological Development Zone, Daxing District, Beijing Patentee after: Sinovision Technology (Beijing) Co.,Ltd. Address before: 100176 floor 1, Yuehong building, No.13, Yongchang North Road, Beijing Economic and Technological Development Zone, Daxing District, Beijing Patentee before: SAINUO WEISHENG TECHNOLOGY (BEIJING) Co.,Ltd. |
|
| CP03 | Change of name, title or address | ||
| PC01 | Cancellation of the registration of the contract for pledge of patent right |
Date of cancellation: 20220406 Granted publication date: 20180420 Pledgee: Beijing Zhongguancun bank Limited by Share Ltd. Pledgor: SAINUO WEISHENG TECHNOLOGY (BEIJING) Co.,Ltd. Registration number: Y2020990000200 |
|
| PC01 | Cancellation of the registration of the contract for pledge of patent right |