CN104182932B - CT (Computed Tomography) device, CT image system and CT image generation method - Google Patents
CT (Computed Tomography) device, CT image system and CT image generation method Download PDFInfo
- Publication number
- CN104182932B CN104182932B CN201310201095.9A CN201310201095A CN104182932B CN 104182932 B CN104182932 B CN 104182932B CN 201310201095 A CN201310201095 A CN 201310201095A CN 104182932 B CN104182932 B CN 104182932B
- Authority
- CN
- China
- Prior art keywords
- image
- reference substance
- iteration
- images
- scanning area
- 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.)
- Expired - Fee Related
Links
Landscapes
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention provides a CT (Computed Tomography) device, a CT image system and a CT image generation method. On the premise that safety is guaranteed, the CT image quality and the CT image generation efficiency are both considered. The CT device scans a scanning area through X-rays to generate a CT image of a scanned object in the scanning area, and is provided with a reference substance device and a CT image generation device, wherein the reference substance device is arranged in a specified position in the scanning area; and according to the known CT image information of the reference substance device and scanning data, which is obtained by scanning, of the scanning area, the CT image of the scanned object is generated. When the CT image of the scanned object is generated by utilizing an iterative reconstruction way, the iterative step length can be determined especially by utilizing the known CT image information of the reference substance device as well as a reconstruction image and an updating image corresponding to the reference substance device in iteration. Therefore, iteration frequencies in iterative reconstruction can be effectively reduced, and the iterative reconstruction efficiency is improved.
Description
Technical field
The present invention relates to CT devices, CT picture systems and CT image generating methods, more particularly, to improve CT images and change
In generation, rebuilds CT devices, CT picture systems and the CT image generating methods of efficiency.
Background technology
X ray computer fault imaging（CT）Technology has been widely used for checking human body that CT images are used as to disease
The foundation of disease diagnosis has the history of 30 years, and CT Image Reconstruction Technology is studied to reduce radiation dose, improve CT images
Quality, the hot issue reduced during image artifacts are always studied and be clinical.
In practical application, CT Image Reconstruction Technology mainly includes filtered back projection's mode and iterative approximation mode.Wherein, filter
Ripple back projection mode is the traditional approach of CT image reconstructions, is widely used in current CT products.But
In filtered back projection's mode, the data for projection of reconstruction image is assumed to be noiseless interference, and in fact, noise is accompanied by
What data for projection was present all the time, it is especially even more so in the case of low-dose scanning, therefore, it is difficult to obtaining high-quality CT
Image.However as the development of clinic diagnosis, the breadth and depth of CT clinical practices has all day by day reached unprecedented height,
Under this new situation background, the security consideration that industry is used CT has new, higher requirement with picture quality.
This just causes filtered back projection's mode to be difficult to meet new demand.Even if in low and middle-end application, filtered back projection's mode is still
Need new more accurate back projection method to reduce artifact, improve picture quality.
The new demand for more than, in high-end applications, iterative approximation mode is taken seriously and studies.Iterative approximation mode can
With image artifacts caused by processing electronic noise and other physical factor institutes well, so as in the situation for ensureing picture quality
Under, reduce x-ray dose when checking.In the past, cannot actually face because its huge amount of calculation causes image taking speed slow
Bed application.In recent years, developing rapidly with computer hardware and computational science, iterative approximation mode be applied to actual product into
In order to may, and the pay attention to day by day with society to medical treatment ＆ health, impact of the X-radiation in CT diagnosis to health
Increasingly paid close attention to by people, low X-ray radiation dosage has become the future trend of CT development.Therefore iterative approximation mode
More and more widely paid close attention to, be current study hotspot.
Iterative reconstruction process mainly includes the projection of multiple loop iteration and back projection's process, in this process, passes through
The iteration of each wheel, the image meeting Step wise approximation ideal image for obtaining, can ensure good image in the case of low noise
Resolution and definition, but iteration many times is generally required, and time-consuming, becomes a Main Bottleneck of clinical practice.
Therefore, in current CT fields, all there is respective skill in filtered back projection's mode and iterative approximation mode
Art problem, in the urgent need on the premise of safety is guaranteed, taking into account CT picture qualities and CT image formation efficiencies.
The content of the invention
Based on background above, it is an object of the present invention to provide a kind of CT devices, CT picture systems and CT images generation side
Method, can take into account CT picture qualities and CT image formation efficiencies on the premise of safety is guaranteed.
In order to achieve the above object, it is raw the present invention relates to a kind of CT devices, are scanned by X-ray to scanning area
Into the CT images of the sweep object in the scanning area, it is characterised in that possess：Reference substance device, is arranged on described
Assigned position in scanning area；And CT video generation devices, known CT image informations according to the reference substance device,
And by scanning the scan data of the scanning area for obtaining, generate the CT images of the sweep object.
CT devices of the invention, by sweep object（Such as human body）Assigned position in neighbouring scanning area
Increase known to CT image informations（For example give material）Reference substance device, and generate sweep object CT images when using ginseng
CT image informations known to thing device are examined, can be taken into account CT picture qualities and CT images are generated on the premise of safety is guaranteed
Efficiency.
Can also be that the CT video generation devices utilize the known CT of the reference substance device in above-mentioned CT devices
Image information, the scan data of the scanning area to being obtained by scanning is iterated reconstruction, thus generates the scanning
The CT images of object.
Thus, in iterative approximation mode, CT image informations are generating sweep object with reference to known to reference substance device
CT images, can efficiently reduce the number of times of iteration in iterative approximation, improve the efficiency of iterative approximation.Due to iterative approximation mode
With safe and picture quality it is high the characteristics of, here further improves the efficiency of iterative approximation mode, so as to significantly
Improve the practicality of iterative approximation mode.
Can also be that in the iterative approximation, the CT video generation devices are according to the ginseng in above-mentioned CT devices
Examine the current of the reference object area that reference substance device described in the known CT images and the scanning area of thing device is located
Reconstruction image and more new images, determine current iteration used in step-length.
In iterative approximation, the size of step-length arranges the convergence rate to iterative approximation and has a very big impact.By profit
With the known CT images of reference substance device and the reconstruction image and more new images in iteration corresponding to reference substance device come
Iteration step length is adaptively determined, step-length can be more suitably determined, so as to greatly accelerate the process of iteration convergence, raising changes
For the efficiency of reconstruction mode.
Can also be in above-mentioned CT devices, in the iterative approximation, the scanning area scan data with it is described
Less than in the case of the first defined threshold, the CT schemes difference between the data for projection of the current reconstruction image of scanning area
As arbitrary width of the generating means used in current iteration in prescribed limit.
Can also be that in the iterative approximation, the step-length used in last iteration is less than the in above-mentioned CT devices
In the case of two defined thresholds, arbitrary width of the CT video generation devices used in current iteration in prescribed limit.
Here, the local optimum for being likely to occur during iteration step length is determined using reference substance device is asked
Topic, by using the arbitrary width in prescribed limit local optimum is jumped out, and prevents from vibrating and restraining in the range of local optimum
It is slow or can not restrain, so as to improve the precision and efficiency of iteration convergence.
Can also be in above-mentioned CT devices, in the iterative approximation, in the current reconstruction of the reference object area
In the case that image reaches definite quality, or the quality of the reconstruction image of the reference object area is less than in current iteration
In the case of in secondary iteration, the CT video generation devices stop iteration.
In the prior art, stopping criterion for iteration cannot be met all the time sometimes, causes iterative approximation convergence judgement mistake
The problem of effect.In this regard, setting stopping criterion for iteration based on reconstruction image corresponding with reference substance device, prior art is can solve the problem that
In the problems referred to above, be accurately and reliably iterated reconstruction convergence judge.
Can also be in above-mentioned CT devices, in known CT images and the reference object area of the reference substance device
Current reconstruction image between difference less than in the case of the 3rd defined threshold, the CT video generation devices are judged as institute
State the current reconstruction image with reference to object area and reach definite quality, the reference substance device known CT images with it is described
In the case of being more than in once iteration in current iteration with reference to the difference between the reconstruction image of object area, the CT images
Generating means are judged as that the quality of the reconstruction image of the reference object area is less than in once iteration in current iteration.
Here, there is provided the specific stopping criterion for iteration based on reference substance device.By according to above iteration ends bar
Part, can accurately and reliably be iterated reconstruction convergence and judge.
Can also be that in the iterative approximation, the CT video generation devices also carry out canonical in above-mentioned CT devices
Change Filtering Processing, the regularization Filtering Processing is the reconstruction image of the scan data to the scanning area and the scanning area
Data for projection between difference corresponding to residual image carry out the process of regularization filtering, in the regularization Filtering Processing
In, the CT video generation devices according to the noise of the reconstruction image of the reference object area, used in determining regularization filtering
Filter parameter.
In the prior art, when carrying out regularization Filtering Processing in iterative approximation, set according to the noise intensity of image
The parameter of wave filter is put, but because the noise intensity for obtaining has error, leads to not the parameter of the appropriate wave filter of setting.It is right
This, by using the noise of reconstruction image corresponding with reference substance device, can solve the problem that the problems of the prior art, in regularization
The parameter of wave filter is set adaptively in Filtering Processing.
Can also be to be disposed around the scanning to reference substance device one or split in above-mentioned CT devices
In the annular region of object, in the iterative approximation, the CT video generation devices by initial pictures than the annuluses
The pixel value in the more outward region in domain is set to 0.
When reference substance device is arranged with annular around sweep object, it is believed that in the region in the outside of the annular region
There is no sweep object.Therefore, by the way that the pixel value in the region in the outside of the annular region is set to into 0, initial graph is enabled to
As closer to final result image, so as to the process for accelerating to restrain, improving the efficiency of iterative approximation.
Can also be many of annular, the rectangle of one, split that the reference substance device is integrated in above-mentioned CT devices
A certain kind in individual rectangle, multiple circles of split, evenly around sweep object configuration.
Here, specifically listing several preferred configuration of reference substance device.Thereby, it is possible to more easily grasp ginseng
Allocation position of the thing device in scanning area is examined, so as to preferably grasp the CT image informations of reference substance device.
In addition, in order to reach the purpose of the present invention, the invention further relates to a kind of CT picture systems, it is characterised in that possess
The CT images output of the CT images of the sweep object that above-mentioned CT devices and output are generated by the CT video generation devices
Device.
CT picture systems of the invention, by sweep object（Such as human body）Regulation in neighbouring scanning area
Position increases known to CT image informations（For example give material）Reference substance device, and generate sweep object CT images when profit
With CT image informations known to reference substance device, CT picture qualities and CT images can be taken into account on the premise of safety is guaranteed
Formation efficiency.Thereby, it is possible to export at faster speed（For example show）The higher CT images of quality.
In addition, in order to reach the purpose of the present invention, the invention further relates to a kind of CT image generating methods, according to being penetrated by X
Line scanning area is scanned obtained from scan data, generate the CT figures of the sweep object in the scanning area
Picture, it is characterised in that using the known CT images of the reference substance device at the assigned position being arranged in the scanning area
Information, to the scan data reconstruction is iterated, and thus generates the CT images of the sweep object；In the iterative approximation
In, so that image obtained from back projection is carried out to scan data as initial reconstruction image, following step is performed repeatedly（1）~
（3）Until meeting iteration stopping condition：（1）The reconstruction image of the scan data based on the scanning area and the scanning area
Data for projection between difference, obtain the more new images of the scanning area；（2）According to known to the reference substance device
The reconstruction image and more new images of the reference object area that reference substance device described in CT images and the scanning area is located,
Determine the step-length used in current iteration；（3）According to the reconstruction image and more new images of the scanning area, utilize what is determined
The step-length, obtains the new reconstruction image of the scanning area；When iteration stopping condition is met, according to the scanning area
Current reconstruction image, generate sweep object CT images.
CT image generating methods of the invention, based on iterative approximation mode sweep object is being generated（Such as human body）'s
During CT images, using known to the CT image informations at the assigned position being arranged in the scanning area near sweep object（For example
Given material）Reference substance device, the reference substance device especially using the known CT images of reference substance device and in iteration
Corresponding reconstruction image and more new images are from adaptively determining iteration step length.Thereby, it is possible to efficiently reduce iterative approximation
The number of times of middle iteration, improves the efficiency of iterative approximation.Because iterative approximation mode has the safe and high spy of picture quality
Point, here further improves the efficiency of iterative approximation mode, so as to substantially increase the practicality of iterative approximation mode.
CT devices of the invention, CT picture systems and CT image generating methods, can guarantee the premise of safety
Under, take into account CT picture qualities and CT image formation efficiencies.Wherein, the present invention and it is defined in mode listed above.For example, at this
In the CT image generating methods of invention, the above-mentioned adaptive step determining method based on reference substance device can not only be adopted,
But also can individually adopt or appropriately combined above-mentioned arbitrary width establishing method, the iterative approximation receipts based on reference substance device
Hold back decision method, adaptive regularization filtering parameter method to set up, the iteration based on reference substance device based on reference substance device
Initialisation image modification method etc..And, each step in the CT image generating methods of the present invention is also used as functional module
Realize.
Description of the drawings
Fig. 1 is the structure chart of CT picture systems.
Fig. 2A is the schematic diagram of of the allocation position of reference substance device.
Fig. 2 B are the schematic diagrams of several configurations of reference substance device.
Fig. 3 is the basic step figure of iterative approximation.
Fig. 4 A are iterative process schematic diagrams in the past using step-length during fixed step size in the case of larger.
Fig. 4 B are iterative process schematic diagrams in the past using step-length during fixed step size in the case of less.
Fig. 4 C are using the iterative process schematic diagram in the case of adaptive step.
Fig. 4 D are the schematic diagrams that adaptive step is calculated based on reference substance device.
Fig. 4 E be fixed step size with adaptive step in the case of experimental result comparison diagram.
Fig. 5 be adaptive step with self adaptation+arbitrary width in the case of experimental result comparison diagram.
Fig. 6 is the schematic diagram that the iterative approximation convergence based on reference substance device judges.
Fig. 7 is the schematic diagram that the adaptive regularization filtering parameter based on reference substance device is arranged.
Iteration initialization image correction of Fig. 8 descriptions based on reference substance device is processed.
Fig. 9 is the flow chart of the CT image generating methods for representing the present invention.
Figure 10 is the flow chart of of the CT image generating methods for representing the present invention.
Specific embodiment
First, the structure of CT picture systems involved in the present invention is illustrated.Fig. 1 is the structure chart of CT picture systems.Such as Fig. 1
Shown, CT picture systems mainly include CT devices 1 and CT image output devices 2.CT devices 1 are for example swept using existing X-ray
Device is retouched, scanning area is scanned by X-ray, generate the CT images of the sweep object in scanning area.Here, sweeping
It is, for example, human body etc. to retouch object.CT image output devices 2 export the CT images of the sweep object generated by CT devices 1.Here, CT
Image output device 2 is typically CT image display devices, and the CT figures of the sweep object generated by CT devices 1 are shown on screen
Picture.Certainly, CT image output devices are not limited to CT image display devices, it is also possible to send what is generated by CT devices 1 by network
Printer of CT images that the data transmission interface of CT images, printing are generated by CT devices 1 etc..
The characteristic structural of CT devices 1 involved in the present invention mainly includes reference substance device 11 and CT video generation devices
12.Reference substance device 11 is arranged on the assigned position in scanning area, remains to describe in detail hereinafter with regard to reference substance device 11.CT images
Generating means 12 are for example realized by general processor or special integrated circuit, schemed according to the known CT of reference substance device 11
Scan data as information and by scanning the scanning area for obtaining, generates the CT images of sweep object.
Hereinafter, reference substance device 11 proposed by the present invention is illustrated.The present invention has increased reference newly in CT picture systems
Thing device 11.Fig. 2A is the schematic diagram of of the allocation position of reference substance device.As shown in Figure 2 A, in the existing of CT devices 1
X-ray scanning device in, swing-around trajectory 201 is the swing-around trajectory of x-ray source 202 and detector 203.As reference substance device 11
One, reference substance 204 be arranged on as the body scans region 205 of sweep object around.Can be by any herein with reference to thing
Solid-state high-purity material is constituted, such as by nonmetallic materials such as silicon, synthesize the organic materials such as macromolecular material or ferrum, copper etc.
Metal material etc. constitute, and the purity and concordance of material the higher the better, so contribute to the corresponding CT image values of reference substance
It is a constant, the CT image informations such as corresponding such as CT image values of the reference substance that each material manufacture goes out can in advance test survey
Examination is obtained, and is known.Fig. 2 B are the schematic diagrams of several configurations of reference substance device.As shown in Figure 2 B, it is distributed in human body
Scanning area（The elliptic region of figure immediate vicinity）The reference substance 204 of surrounding can be the annular of one, the rectangle of one, split
Multiple rectangles, split it is multiple circle in a certain kind, match somebody with somebody evenly around the body scans region as sweep object
Put.Here, the shape of reference substance 204 is not limited, distributing position is not limited, but needs to know position of the reference substance 204 in scanning area
Put, so as to obtain the pixel region corresponding to reference substance in CT images（Also referred to as refer to object area）.Here, it is preferred that reference substance is equal
It is distributed in evenly around the body scans region as sweep object, will be illustrated as schematic diagram using annular in present specification.
Used as an embodiment of the invention, CT video generation devices 12 are realized by processor, mainly included for base
The general processor units of control such as this program, parameter and it is exclusively used in the iterative approximation processing unit of iterative approximation.CT images
Generating means 12 utilize the known CT image informations of reference substance device 11, the scanning number of the scanning area to being obtained by scanning
According to reconstruction is iterated, the CT images of sweep object are thus generated.
Fig. 3 is the basic step figure of iterative approximation.Below it is briefly described, CT devices 1 for example utilize basic X-ray
Scanning device is actually scanned to scanning area（Step 301）Actual scanning data S are obtained, actual scanning data S are after filtering
Back projection's process（Step 309）Obtain the initial pictures of iterative approximationThe initial pictures are projected（Step 308）Obtain
Calculate data for projectionThen to S andCarry out mathematic interpolation（Step 302）To projection residual errors Δ S, projection residual errors Δ S is entered again
Row back projection（Step 303）, residual image Δ I is obtained, regularization filtering is carried out to residual image Δ I（Step 304）Obtain more
New images Δ UI, then judges whether Δ UI all pixels are approximately 0（Step 306）, if not being approximately 0, then adding up updates
（Step 307）Reconstruction image, cumulative renewal is that more new images Δ UI is weighted into a step-length（Also referred to as relaxation factor）Add up after α
In the reconstruction image obtained to last round of loop iteration, i.e.,
（Formula 1）
The initial reconstructed image of wherein first run iteration isAgain to reconstruction imageCarry out projection to obtain calculating data for projection
The iterative process of next round is carried out, until more new images Δ UI all pixels are approximately 0 end, reconstruction image at this momentAs
The final reconstruction image of whole iterative approximation.
In iterative approximation, the process of iteration may be considered the process of object function J minimums, cause to solve I
Object function J is minimized, i.e.,
（Formula 2）
Wherein A is sytem matrix, and I is CT images, and S is actual scanning data, and P (I) is image prior item of information, by iteration
During regularization filtering embody.The object function for solving above is minimized, typically using gradient descent method, so as to obtain
（Formula 3）
Step-length in the as cumulative renewal processes of wherein α（Relaxation factor）.The step-length（Relaxation factor）Size arrange it is right
Iterative approximation convergence rate has a very big impact, detailed description below.
Fig. 4 A are iterative process schematic diagrams in the past using step-length during fixed step size in the case of larger.As shown in Figure 4 A,
If arranging larger step-length（Relaxation factor）α, when close minima I of the value of object function J∞Or run into local minimum
When, renewal process can cause target function value to oscillate around in minima or local minimum, even if iteration many times can not
Minima I is converged to quickly∞。
Fig. 4 B are iterative process schematic diagrams in the past using step-length during fixed step size in the case of less.As shown in Figure 4 B,
When the less step-length of setting（Relaxation factor）During α, each time iteration update all seldom, iterative process is very slow, needs also exist for
Iteration many times can just converge to minima I∞。
If it is possible to according to the quality of reconstruction image in iterative process come adaptively adjusting step（Relaxation factor）
The size of α, the process of iteration convergence will greatly be accelerated.Fig. 4 C are using the iterative process in the case of adaptive step
Schematic diagram.As shown in Figure 4 C, compared with the situation of Fig. 4 A and Fig. 4 B, the process of iteration convergence is greatly accelerated.
In order to according to the quality of reconstruction image in iterative process come adaptively adjusting step（Relaxation factor）The size of α,
Adaptive step-length is calculated in the present invention according to reference substance device 11（Relaxation factor）.In iterative approximation, CT images are generated
The reference substance area that device 12 is located according to reference substance device 11 in the known CT images and scanning area of reference substance device 11
The current reconstruction image and more new images in domain, determines the step-length used in current iteration.As a concrete example, such as Fig. 4 D institutes
Show, for the i-th given wheel iteration, select a step-length（Relaxation factor）α so that in this step-length（Relaxation factor）Under α, weight
Build minimum with reference to the squared difference of object area CT image values corresponding with reference substance material after cumulative renewal in image, Ke Yiyong
Formula is expressed as
（Formula 4）
Wherein IrFor the known CT images of reference substance device 11, reference substance image-region is represented, its value is constant, that is, join
The corresponding CT image values of thing material are examined,For the reconstruction image in current reference object image region, Δ UIrFor current reference object image
The more new images in region.
Fig. 4 E be fixed step size with adaptive step in the case of experimental result comparison diagram.In the emulation shown in Fig. 4 E
In experiment, stopping criterion for iteration is set to residual error less than 50, it can be seen that use adaptive step（Relaxation factor）Method can be with
Effectively improve the convergence rate of iteration.
As described above, by using the known CT images of reference substance device 11 and in iteration the institute of reference substance device 11
Corresponding reconstruction image and more new images can more suitably determine step-length, so as to greatly accelerate repeatedly determining iteration step length
The process held back is withheld, the efficiency of iterative approximation mode is improved.
Hereinafter, the variation that the above-mentioned adaptive step based on reference substance device is determined is illustrated.It is above-mentioned from
In adapting to the decision process of step-length, reference substance area image is the regional area in whole CT images, therefore, based on the region
Optimized step-length（Relaxation factor）Iterative process may be caused to be absorbed in local optimum, although having rebuild with reference to object area
Very high precision is reached, but whole image reconstructed results are still not up to optimum, so as to cause stopping iteration or iteration slow.
In this case, local optimum can effectively be jumped out using random step-length.
As a kind of situation that local optimum is jumped out using random step-length, in iterative approximation when residual error very little but not yet
When meeting stopping criterion for iteration, using random step-length local optimum is jumped out.That is, in the scan data and scanning area of scanning area
Difference between the data for projection of the current reconstruction image in domain（That is residual error）In the case of the first defined threshold T1, CT figures
As arbitrary width of the generating means 12 used in current iteration in prescribed limit.Here, above-mentioned residual error can be residual using projection
Difference Δ S, it is also possible to using residual image Δ I.First defined threshold T1 typically could be arranged to the 5% of first run iteration residual error or
Determine according to situation test is embodied as.In addition, arbitrary width is to be controlled in a range of random number, according to actual reality
Fixed this random number range of test.
As another kind of situation that local optimum is jumped out using random step-length, step-length very little but residual error in iterative approximation
When not yet meeting stopping criterion for iteration, using random step-length local optimum is jumped out.That is, the step-length used in last iteration is little
In the case of the second defined threshold, arbitrary width of the CT video generation devices 11 used in current iteration in prescribed limit.
Here, the second defined threshold typically could be arranged to first run step-length（Relaxation factor）5% or according to be embodied as situation test
Determine.In addition, arbitrary width is to be controlled in a range of random number, this random number model is determined according to actual experiment
Enclose.
Fig. 5 be adaptive step with self adaptation+arbitrary width in the case of experimental result comparison diagram.As shown in figure 5,
In the case of low residual error（Corresponding to above-mentioned the first situation）, use in the case where the convergence of adaptive step successive iterations is relatively slow
Arbitrary width enables to the iteration ends target that iteration soon converges to setting（Meet stopping criterion for iteration）.Equally, exist
In the case of low step-length（Corresponding to above-mentioned second case）, it is also possible to obtain similar effect.That is, for being filled using reference substance
The problem of the local optimum being likely to occur during putting 11 decision iteration step lengths, by using random step in the scenario above
Length jumps out local optimum, is prevented from vibrating and restraining slow or can not restrain in the range of local optimum, so as to improve iteration
The precision and efficiency of convergence.
Hereinafter, another embodiment of the invention is illustrated, i.e., reconstruction convergence is iterated based on reference substance device
Judge.
In traditional iterative approximation, as described in Fig. 3, often whether it is approximately 0 according to more new images and is used as iteration
End condition, or be used as stopping criterion for iteration according to whether projection residual errors are approximately 0, but due to projection and back projection's mould
Type is that approximate model, more new images and projection residual errors may can not be approximately all the time 0, causes method sometimes can fail.Cause
This present invention utilizes reference substance device 11, it is proposed that a kind of convergence decision procedure based on reference substance device.That is, in iterative approximation
In, in the case where the current reconstruction image with reference to object area reaches definite quality, or with reference to the reconstruction image of object area
Quality in current iteration less than in the case of in last iteration, CT video generation devices 12 stop iteration.
As the example of the specific stopping criterion for iteration based on reference substance device, in the known CT of reference substance device 11
Image and with reference to the difference between the current reconstruction image of object area less than in the case of the 3rd defined threshold, CT images are generated
Device 12 is judged as reaching definite quality with reference to the current reconstruction image of object area.In addition, known to reference substance device 11
CT images and in the case of being more than in once iteration in current iteration with reference to the difference between the reconstruction image of object area,
CT video generation devices 12 are judged as that the quality of the reconstruction image with reference to object area is less than in once iteration in current iteration.
Below control illustrates the example of the above-mentioned specific stopping criterion for iteration based on reference substance device.Fig. 6 is based on reference substance
The schematic diagram that the iterative approximation convergence of device judges.As shown in fig. 6, working as in reconstruction image with reference to object area and reference substance material pair
The squared difference of the CT image values answered is less than threshold value TstopWhen, or the squared difference is more than in last iteration in current iteration
The squared difference, then stop iteration, can be expressed as with formula
Or
（Formula 5）
Wherein TstopCan be specified by user's image quality level as requested.Here, reference substance device 11 is known
CT images and be not limited to show as in reconstruction image with reference to object area and ginseng with reference to the difference between the reconstruction image of object area
Examine the squared difference of the corresponding CT image values of thing material, it is also possible to showed with other appropriate values such as the absolute values of the difference,
Now it is appropriately arranged with corresponding threshold value.
In the above-mentioned iterative approximation convergence based on reference substance device judges, because the CT images of reference substance device 11 are
Know, therefore compared with the situation using approximate model such as more new images or residual error is based in the past, can be more accurately and reliably
It is iterated reconstruction convergence to judge.
Hereinafter, another embodiment of the invention is illustrated, i.e., canonical is arranged based on reference substance device-adaptive
Change filtering parameter.
Such as in iterative approximation, regularization is filtered corresponding to the prior information item in object function, adjacent picture in image
Element often have this kind of prior information such as approximate pixel value by regularization filtering to be fused to object function in, it is past
Toward the complex filters kept using some smoothing filters or edge, such as Gaussian smoothing filter, bilateral filtering, Geman filters
Ripple etc., the parameter in these wave filter is often being configured according to the noise intensity of image.But in an iterative process, often
The noise intensity of one wheel iteration result is different, in practical application, typically arrange a preset parameter or according in image certain
The preferable part of concordance, such as certain organ-tissue region estimates noise intensity and is adjusted, but the organ in reconstruction image
Tissue regions can not possibly have crash consistency, estimate that the noise intensity for obtaining there is also certain error.
For the problems referred to above of the prior art, a kind of adaptive regularization based on reference substance device is herein proposed and has filtered
Wave parameter set-up mode.That is, CT video generation devices 12 also carry out regularization Filtering Processing, to the scan data of scanning area with
The residual image corresponding to difference between the data for projection of the reconstruction image of scanning area carries out the process of regularization filtering,
In regularization Filtering Processing, CT video generation devices 12 determine regularization filter according to the noise of the reconstruction image with reference to object area
Filter parameter used in ripple.Fig. 7 is the schematic diagram that the adaptive regularization filtering parameter based on reference substance device is arranged.Such as
Shown in Fig. 7, standard deviation SD of the noise that object area is referred in reconstruction image is calculatednoise（Step 701）, then according to this standard deviation
Go the parameter that regularization wave filter is set（Step 702）, the different wave filter of design parameter installation warrants and it is different, for example for
Gaussian smoothing filter, the variance in Gaussian filter is proportional to standard deviation SDnoise, i.e., stronger noise needs larger smooth strong
Degree.SDnoiseComputational methods such as following formula：
（Formula 6）
Wherein, mean is mean value function,For reference substance area pixel in image, N is reference substance area in all images
Domain sum of all pixels.
In the above-mentioned embodiment that regularization filtering parameter is arranged based on reference substance device-adaptive, using known
Reference substance device 11 arranging regularization filtering parameter, compared with being arranged based on experience estimation in prior art, can be more
Regularization filtering parameter is suitably set.Particularly reference substance device 11 have it is very high it is conforming in the case of, can be with
High accuracy arranges regularization filtering parameter.
Hereinafter, another embodiment of the invention is illustrated, i.e., based on reference substance device amendment iteration initialization figure
Picture.
Iteration initialization image correction of Fig. 8 descriptions based on reference substance device is processed.In iterative approximation, initial pictures are past
It is past to use traditional filtered back projection's result images, but when actual projection data is incomplete, as shown in the left figure in Fig. 8,
The form and scanning thing of the filtered backprojection image for obtaining is by with very big difference.
It is the optimization process of object function in view of iterative process, if the image initial state of iteration nearer it is to finally
Image, then can Accelerated iteration process.Therefore, it is disposed around sweep object in the one of reference substance device 11 or split
Annular region in the case of, in iterative approximation, CT video generation devices 12 by initial pictures than the annular region more
The pixel value in region in the outer part is set to 0.That is, as shown in figure 8, the region 801 outside reference substance device 802, due to being all empty
Gas, the value of final result image must be all 0, therefore, in initialisation image, the picture in the region 801 outside reference substance device is set
Element value is 0.Initial pictures can so be caused closer to final result image, so as to the process for accelerating to restrain.
Hereinafter, CT image generating methods involved in the present invention are illustrated.CT images generation side involved in the present invention
Method can be performed by CT devices 1, more specifically can be performed by the CT video generation devices 12 of CT devices 1.Fig. 9 is to represent the present invention
CT image generating methods flow chart.As shown in figure 9, CT image generating methods involved in the present invention are according to by X-ray
Scan data obtained from being scanned to scanning area, generates the CT images of the sweep object in scanning area.Wherein,
It is right using the known CT image informations of the reference substance device 12 at the assigned position being arranged in scanning area in CT devices 1
Scan data is iterated reconstruction, thus generates the CT images of sweep object.
In iterative approximation, first in step sl, so that image obtained from back projection is carried out to scan data as first
The reconstruction image of beginning.Then, in step s 2, the projection of the reconstruction image of the scan data based on scanning area and scanning area
Difference between data, obtains the more new images of scanning area.In step s3, schemed according to the known CT of reference substance device 11
The reconstruction image and more new images of the reference object area that reference substance device 11 is located, determines that this changes in picture and scanning area
Step-length used in generation.In step s 4, according to the reconstruction image and more new images of scanning area in last iteration, using being determined
Fixed step-length, obtains the new reconstruction image of scanning area.In step s 5, judge whether to meet stopping criterion for iteration, if
Stopping criterion for iteration is unsatisfactory for, then execution step S2～S4 repeatedly.When iteration stopping condition is met, working as according to scanning area
Front reconstruction image, generates the CT images of sweep object and terminates.
CT image generating methods of the invention, using iterative approximation mode sweep object is being generated（Such as human body）'s
During CT images, using known to the CT image informations at the assigned position being arranged in the scanning area near sweep object（For example
Given material）Reference substance device, the reference substance device especially using the known CT images of reference substance device and in iteration
Corresponding reconstruction image and more new images are determining iteration step length.Iteration in thereby, it is possible to efficiently reduce iterative approximation
Number of times, improves the efficiency of iterative approximation.The characteristics of having safe and picture quality high due to iterative approximation mode, here is again
The efficiency of iterative approximation mode is further increased, so as to substantially increase the practicality of iterative approximation mode.
In the CT image generating methods of the present invention, can not only adopt in above-mentioned embodiment based on reference substance device
Adaptive step determining method, but also can individually adopt or the above-mentioned variation of appropriate combination in arbitrary width setting side
In method, above-mentioned other embodiment based on reference substance device iterative approximation convergence decision method, based on reference substance device
Adaptive regularization filtering parameter method to set up, iteration initialization image correcting method based on reference substance device etc..Hereinafter, have
Body explanation combines of the respective embodiments described above and the CT image generating methods of the invention after variation.
Figure 10 is the flow chart of of the CT image generating methods for representing the present invention.First, in step 901, utilize
The iteration initialization image correcting method based on reference substance device in above-mentioned embodiment is to filtered back projection's initialisation image
It is modified.Then in step 902, carry out projection and obtain data for projection.In step 903, projection residual errors are further calculated.
In step 904, then back projection is carried out to projection residual errors and obtains residual image.In step 905, using in above-mentioned embodiment
The adaptive regularization filtering parameter arranged based on reference substance device is carried out regularization filtering to residual image and obtains more new images.
Then in step 906, the cumulative of projection residual errors is judged and whether less than threshold value T1.If being less than, in step 908 using above-mentioned
Random adaptive step in variation（Relaxation factor）.Otherwise, it is based on reference substance according to above-mentioned embodiment in step 907
Device estimation self-adaptive step-length.Then in step 909, to the random adaptive step obtained in more new images step 908
Or the adaptive step weighting obtained in step 907, and cumulative renewal arrives reconstruction image.In step 910, current reconstruction is calculated
The noise mean square value and standard deviation of the reference object area respective pixel in image.Here, the standard deviation calculated in step 910 is used for
The parameter of regularization filtering in Automatic adjusument next round iteration, the mean-square value calculated in step 910 is used to carry out the condition of convergence
Judgement.In step 911, if meeting the stopping criterion for iteration based on reference substance device in above-mentioned embodiment, stop
Iteration obtains final result, and otherwise continuing back to step 902 carries out next step iteration.
In above-mentioned of CT image generating methods, it is of course possible to according to the CT devices 1 having been described above in this specification
Each embodiment and variation further deforming.For example, in step 906, it is also possible to judge that step-length is as previously mentioned
It is no to be less than the second defined threshold, the step-length very little but when residual error not yet meets stopping criterion for iteration in iterative approximation, using random
Step-length jump out local optimum.In addition, in step 911, rule can be reached according to the current reconstruction image with reference to object area
Determine quality or be less than in last iteration to judge to meet in current iteration according to the quality of the reconstruction image with reference to object area
The condition of convergence.
Above by reference to the specific embodiment for having illustrated the present invention.Wherein, specific embodiment described above is only
It is the specific example of the present invention, for understanding the present invention, rather than limits the scope of the present invention.Those skilled in the art can
Various modifications, combination and the reasonable omission of key element are carried out to specific embodiment based on the technological thought of the present invention, is thus obtained
Mode be intended to be included within the scope of the present invention.
Claims (10)
1. a kind of CT devices, are scanned by X-ray to scanning area, generate the sweep object in the scanning area
CT images, it is characterised in that possess：
Reference substance device, the assigned position being arranged in the scanning area；And
CT video generation devices, according to the known CT image informations of the reference substance device and by scanning the institute for obtaining
The scan data of scanning area is stated, the CT images of the sweep object are generated,
The CT video generation devices utilize the known CT image informations of the reference substance device, to the institute obtained by scanning
The scan data for stating scanning area is iterated reconstruction, thus generates the CT images of the sweep object,
In the iterative approximation, the CT video generation devices according to the known CT images of the reference substance device and
The current reconstruction image and more new images of the reference object area that reference substance device described in the scanning area is located, determines this
Step-length used in secondary iteration.
2. CT devices as claimed in claim 1, it is characterised in that
In the iterative approximation, in scan data and the current reconstruction image of the scanning area of the scanning area
, less than in the case of the first defined threshold, the CT video generation devices are used in current iteration for difference between data for projection
Arbitrary width in prescribed limit.
3. CT devices as claimed in claim 1, it is characterised in that
In the iterative approximation, less than in the case of the second defined threshold, the CT schemes the step-length used in last iteration
As arbitrary width of the generating means used in current iteration in prescribed limit.
4. CT devices as claimed in claim 1, it is characterised in that
In the iterative approximation, in the case where the current reconstruction image of the reference object area reaches definite quality, or
In the case that quality described in person with reference to the reconstruction image of object area is less than in last iteration in current iteration, the CT images
Generating means stop iteration.
5. CT devices as claimed in claim 4, it is characterised in that
Difference between the known CT images of the reference substance device and the current reconstruction image of the reference object area
In the case of less than the 3rd defined threshold, the CT video generation devices are judged as the current reconstruction figure of the reference object area
As reaching definite quality,
Difference between the known CT images of the reference substance device and the reconstruction image of the reference object area is at this
In the case of being more than in iteration in last iteration, the CT video generation devices are judged as the reconstruction figure of the reference object area
The quality of picture is less than in once iteration in current iteration.
6. CT devices as claimed in claim 1, it is characterised in that
In the iterative approximation, the CT video generation devices also carry out regularization Filtering Processing, the regularization Filtering Processing
Be the scan data to the scanning area and the scanning area reconstruction image data for projection between difference corresponding to
Residual image carry out the process of regularization filtering,
In the regularization Filtering Processing, the CT video generation devices are made an uproar according to the reconstruction image of the reference object area
Sound, determines the filter parameter used in regularization filtering.
7. CT devices as claimed in claim 1, it is characterised in that
It is disposed around in the annular region of the sweep object to reference substance device one or split,
In the iterative approximation, the CT video generation devices are by area more more outward than the annular region in initial pictures
The pixel value in domain is set to 0.
8. CT devices as claimed in claim 1, it is characterised in that
Certain in annular, the rectangle of one, multiple rectangles of split, multiple circles of split that the reference substance device is integrated
One kind, evenly around sweep object configuration.
9. a kind of CT picture systems, it is characterised in that possess：
CT devices any one of claim 1～8；And
CT image output devices, export the CT images of the sweep object generated by the CT video generation devices.
10. a kind of CT image generating methods, the scan data according to obtained from being scanned to scanning area by X-ray, are given birth to
Into the CT images of the sweep object in the scanning area, it is characterised in that
Using the known CT image informations of the reference substance device at the assigned position being arranged in the scanning area, to described
Scan data is iterated reconstruction, thus generates the CT images of the sweep object,
In the iterative approximation, so that image obtained from back projection is carried out to scan data as initial reconstruction image, instead
Following step (1)～(3) are performed again until meeting iteration stopping condition：
(1) difference between scan data based on the scanning area and the data for projection of the reconstruction image of the scanning area
It is different, obtain the more new images of the scanning area；
(2) it is located according to reference substance device described in the known CT images and the scanning area of the reference substance device
Reference object area reconstruction image and more new images, determine current iteration used in step-length；
(3) according to the reconstruction image and more new images of the scanning area, the step-length for being determined is utilized, obtains the scanning
The new reconstruction image in region；
When iteration stopping condition is met, according to the current reconstruction image of the scanning area, the CT figures of sweep object are generated
Picture.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310201095.9A CN104182932B (en) | 2013-05-27 | 2013-05-27 | CT (Computed Tomography) device, CT image system and CT image generation method |
| JP2014108225A JP6247998B2 (en) | 2013-05-27 | 2014-05-26 | CT apparatus, CT image system, and CT image generation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310201095.9A CN104182932B (en) | 2013-05-27 | 2013-05-27 | CT (Computed Tomography) device, CT image system and CT image generation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104182932A CN104182932A (en) | 2014-12-03 |
| CN104182932B true CN104182932B (en) | 2017-04-12 |
Family
ID=51963951
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310201095.9A Expired - Fee Related CN104182932B (en) | 2013-05-27 | 2013-05-27 | CT (Computed Tomography) device, CT image system and CT image generation method |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP6247998B2 (en) |
| CN (1) | CN104182932B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6486100B2 (en) * | 2014-12-22 | 2019-03-20 | キヤノン株式会社 | Image processing apparatus, image processing method, and program |
| JP6595211B2 (en) * | 2015-05-11 | 2019-10-23 | キヤノンメディカルシステムズ株式会社 | Nuclear medicine diagnostic apparatus and nuclear medicine image processing apparatus |
| JP6569733B2 (en) * | 2015-08-17 | 2019-09-04 | 株式会社島津製作所 | Image reconstruction processing method, image reconstruction processing program, and tomographic apparatus equipped with the same |
| JP6699482B2 (en) * | 2016-09-21 | 2020-05-27 | 株式会社島津製作所 | Iterative image reconstruction method, iterative image reconstruction program, and tomography apparatus |
| CN113518221B (en) * | 2016-10-14 | 2024-03-01 | 联发科技股份有限公司 | Video encoding or decoding method and corresponding device |
| CN110613471A (en) * | 2019-10-22 | 2019-12-27 | 合肥工业大学 | CT system and CT image reconstruction method |
| CN114373025A (en) * | 2021-12-06 | 2022-04-19 | 山东师范大学 | CBCT image reconstruction method, system, storage medium and equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101076282A (en) * | 2004-09-30 | 2007-11-21 | 安科锐公司 | Dynamic tracking of moving targets |
| CN101623198A (en) * | 2008-07-08 | 2010-01-13 | 深圳市海博科技有限公司 | Real-time tracking method for dynamic tumor |
| CN101628154A (en) * | 2008-07-16 | 2010-01-20 | 深圳市海博科技有限公司 | Image guiding and tracking method based on prediction |
| CN101897593A (en) * | 2009-05-26 | 2010-12-01 | 清华大学 | Computer chromatography imaging device and method |
| CN102903117A (en) * | 2012-10-24 | 2013-01-30 | 深圳大学 | 3D (three-dimensional) image registration method and device based on conformal geometric algebra |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS5422192A (en) * | 1977-07-21 | 1979-02-19 | Toshiba Corp | Computer tomography device |
| JPS5441689A (en) * | 1977-09-09 | 1979-04-03 | Hitachi Medical Corp | Xxray tomograph |
| JPS6017568A (en) * | 1983-07-11 | 1985-01-29 | Hitachi Ltd | Method and device for processing image |
| US5053958A (en) * | 1988-06-06 | 1991-10-01 | General Electric Company | Method to reduce image reconstruction time in limited-angle ct systems including using initial reconstruction valves for measured projection data during each iteration |
| JP2006014928A (en) * | 2004-07-01 | 2006-01-19 | Fuji Photo Film Co Ltd | Method, device and program for displaying image |
| US20090127451A1 (en) * | 2007-11-16 | 2009-05-21 | Siemens Medical Solutions Usa, Inc. | Devices and Methods for Calibrating Nuclear Medical and Radiological Images |
| US8731269B2 (en) * | 2011-10-19 | 2014-05-20 | Kabushiki Kaisha Toshiba | Method and system for substantially reducing artifacts in circular cone beam computer tomography (CT) |
-
2013
- 2013-05-27 CN CN201310201095.9A patent/CN104182932B/en not_active Expired - Fee Related
-
2014
- 2014-05-26 JP JP2014108225A patent/JP6247998B2/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101076282A (en) * | 2004-09-30 | 2007-11-21 | 安科锐公司 | Dynamic tracking of moving targets |
| CN101623198A (en) * | 2008-07-08 | 2010-01-13 | 深圳市海博科技有限公司 | Real-time tracking method for dynamic tumor |
| CN101628154A (en) * | 2008-07-16 | 2010-01-20 | 深圳市海博科技有限公司 | Image guiding and tracking method based on prediction |
| CN101897593A (en) * | 2009-05-26 | 2010-12-01 | 清华大学 | Computer chromatography imaging device and method |
| CN102903117A (en) * | 2012-10-24 | 2013-01-30 | 深圳大学 | 3D (three-dimensional) image registration method and device based on conformal geometric algebra |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104182932A (en) | 2014-12-03 |
| JP6247998B2 (en) | 2017-12-13 |
| JP2014226562A (en) | 2014-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104182932B (en) | CT (Computed Tomography) device, CT image system and CT image generation method | |
| Kida et al. | Cone beam computed tomography image quality improvement using a deep convolutional neural network | |
| US11508102B2 (en) | Systems and methods for image processing | |
| Zhu et al. | Scatter correction for cone‐beam CT in radiation therapy | |
| Li et al. | 3D tumor localization through real‐time volumetric x‐ray imaging for lung cancer radiotherapy | |
| Harms et al. | Cone‐beam CT‐derived relative stopping power map generation via deep learning for proton radiotherapy | |
| Zeng et al. | Respiratory motion estimation from slowly rotating x‐ray projections: Theory and simulation | |
| Lee et al. | Scatter correction in cone‐beam CT via a half beam blocker technique allowing simultaneous acquisition of scatter and image information | |
| Ouyang et al. | A moving blocker system for cone‐beam computed tomography scatter correction | |
| CN102024251B (en) | System and method for multi-image based virtual non-contrast image enhancement for dual source CT | |
| JP2008528168A (en) | Apparatus and method for correcting or extending X-ray projection | |
| CN104644200A (en) | Method and device for reducing artifacts in computed tomography image reconstruction | |
| US9734574B2 (en) | Image processor, treatment system, and image processing method | |
| Wu et al. | Cone‐beam CT for imaging of the head/brain: development and assessment of scanner prototype and reconstruction algorithms | |
| CN102270349A (en) | Iterative reconstruction of ct images without regularization term | |
| CN112489158B (en) | Enhancement method for low-dose PET image by adopting adaptive network based on cGAN | |
| Sun et al. | Correction for patient table‐induced scattered radiation in cone‐beam computed tomography (CBCT) | |
| Liang et al. | Scatter correction for a clinical cone‐beam CT system using an optimized stationary beam blocker in a single scan | |
| JPWO2017104700A1 (en) | Image processing apparatus and image processing method | |
| Chen et al. | Optimization of the geometry and speed of a moving blocker system for cone‐beam computed tomography scatter correction | |
| Peterlik et al. | Reducing residual‐motion artifacts in iterative 3D CBCT reconstruction in image‐guided radiation therapy | |
| Kim et al. | Convolutional neural network–based metal and streak artifacts reduction in dental CT images with sparse‐view sampling scheme | |
| Fan et al. | Image‐domain shading correction for cone‐beam CT without prior patient information | |
| Cho et al. | Region‐of‐interest image reconstruction with intensity weighting in circular cone‐beam CT for image‐guided radiation therapy | |
| WO2022091869A1 (en) | Medical image processing device, medical image processing method, and program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20160627 Address after: Tokyo, Japan, Japan Applicant after: Hitachi Ltd. Address before: Tokyo, Japan, Japan Applicant before: Hitachi Medical Corporation |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170412 Termination date: 20210527 |