WO2010058329A1

WO2010058329A1 - System and method for x-ray scatter correction

Info

Publication number: WO2010058329A1
Application number: PCT/IB2009/055032
Authority: WO
Inventors: Steffen G. Wiesner; Matthias Bertram; Jens Wiegert; Jan Timmer
Original assignee: Philips Intellectual Property & Standards Gmbh; Koninklijke Philips Electronics N.V.
Priority date: 2008-11-21
Filing date: 2009-11-12
Publication date: 2010-05-27

Abstract

The invention relates to a method of scatter correction of a projection image acquired from an X-ray imaging system. The method is based on estimating an X-ray scatter distribution based on a weighted accumulation of the scatter contribution generated along individual source rays. The method incorporates two weighting functions to account for at least one of the two of a spatial non-uniformity of the X-ray radiation used in the X-ray imaging system and shading effects due to an anti-scatter-grid arranged in the X-ray imaging system.

Description

SYSTEM AND METHOD FOR X-RAY SCATTER CORRECTION

FIELD OF THE INVENTION The invention relates to image processing. More particularly the invention relates to an imaging system for acquiring a digital projection image, a correction unit, a scatter correction method, a computer-readable medium and a program element.

BACKGROUND OF THE INVENTION

The medical profession but also others like material sciences have become to rely more and more on high quality image material such as X-ray images. A radiologist for example is able to adduce a number of clues from X-ray images provided the X-ray images are of sufficient quality. The images acquired from those X-ray imaging systems are projection images. Rays emitted from an X-ray source of the imaging system interact with a target object of which a projection image is to be taken. An intensity of the rays is attenuated in accordance with an x-ray attenuation coefficient of the object at a point of interaction. The rays emitted by the radiation source are incident on a detector after the interaction with the object. The incident rays are then converted into images that can be stored as digital images for further processing or which can be viewed on a monitor. In the medical field the object may be a patient undergoing a chest X-ray whereas in the material sciences the object may be a sample of a work piece that is to be analyzed for micro-cracks. The actual projection image detected at the detector can be thought of as a superposition of a scatter- free primary radiation and a contribution of scattered radiation. The primary radiation comprises the rays that pass through the object undeflected or unscattered whereas the scattered radiation comprises the rays that have been subjected to deflection or scattering. The projection images acquired from C-arm based soft tissue imaging, or cone-beam CT in general, are often corrupted by scatter. Besides deteriorating the quality of the projections and impeding quantitative analysis of the detected intensity values, such scatter contributions to the projections result in even more salient artefacts in volumetric image data reconstructed from the projection images, such as streaks and low frequency inhomogeneities (cupping). The scatter artefacts are a consequence of the scattered radiation. In general, there are at least two factors affecting a distribution of the detected scattered radiation during the acquisition of the projection image

One factor is the spatial non-uniformity of the intensity of the X-ray source in the imaging system. The non-uniformity results, for example, from heel effects, geometrical vignetting or from impacts due to a potential beam shaper. The second factor are shading effects caused by the presence of an anti-scatter-grid (ASG). More specifically these shading effects occur, e.g., when the projections are acquired using a distance between source and detector that differs from the distance the ASG had originally been designed for. This is often the case in, e.g., C-arm imaging systems, especially for volumetric soft-tissue imaging.

A number of correction methods and systems have been put in the place in order to remove the scattered artefacts from the acquired projection image. Such scatter correction methods are for example convolution methods that are used to estimate based on the X-ray projection image a spatial scatter distribution. Those methods rely on the assumption that the spatial scatter distribution can be obtained from the scatter- free primary radiation by convolution filtering of the latter with appropriately chosen kernel functions.

However there is an issue with prior art convolutional scatter estimation methods. Those methods are based on the assumption that the distribution of the scattered radiation is affected by the two factors mentioned above in the same manner or degree as the primary radiation. This assumption normally leads to false estimates of the scattered distribution and thus to lower quality images as the correction is based on such false estimates of the scatter distribution.

There may therefore be a need in the art for correction methods that take due account to the manner and degree in which those two factors affect the distribution of the scattered radiation detected by the detector during the acquisition of the projection image.

There may further be a need for a correction method that can be tailored to an individual imaging system in order to better account for imperfections during the image acquisition process.

SUMMARY OF THE INVENTION

According to one aspect of the present invention there is provided a method for scatter correction for correcting a digital X-ray projection image acquired from an X-ray imaging system. The acquisition of the projection image is based on detected rays of an X-ray radiation source.

According to another aspect of the present invention there is provided a correction unit implementing the scatter correction method according to an exemplary embodiment of the present invention. The correction unit takes as input the acquired projection image. Based on the acquired projection image an X-ray scatter distribution is estimated, wherein a scattered contribution of source rays is weighted with a first weighing image. The estimated scatter distribution is then weighted with a second weighting image. The correction unit then proceeds to correct the projection image by using the re-weighted estimated scatter distribution to compensate for the contribution of scattered radiation intensities in the acquired projection image, e.g., by subtraction. According to an exemplary embodiment of the present invention the X- ray scatter distribution is based on a weighted accumulation of the scatter contribution generated along the source rays of the radiation source. As the scattering occurs along the path taken by the X-ray the weighted accumulation allows taking account of the contribution of each source ray in a more realistic manner. This may yield estimates of higher accuracy.

According to an exemplary embodiment of the present invention, the first weighing image accounts for a spatial variation of the radiation source intensity.

According to another aspect of the present invention, the weighing with the second weighing image accounts for at least one of a normalization with respect to - A -

the spatial variation of the source radiation intensity and a shading effect caused by an anti-scatter grid (ASG) present in the imaging system with which the projection image has been acquired.

According to an exemplary embodiment of the present invention the first and second weighing images are based on first and second gain images acquired from the imaging system. Basing the weighing images on the first and second gain images allows obtaining weighting images tailored to a specific imaging setup of the individual imaging system used as well as imperfections during the image acquisition process. In other words the gain images record a "footprint" of the factors affecting the detected scatter distribution such as a non-uniformity of the radiation intensity of the radiation source and shading effects due to the presence of the ASG in the imaging system. The method according to the invention allows adding value to existing imaging systems by improving image quality without having to go to the expense of installing new equipment other than the correction unit according to the present invention. The first gain image can be obtained by direct exposure of the detector without the object arranged between the radiation source and the detector and without the ASG. The second gain image is obtained similarly but with the ASG in place. The gain images for the purposes of obtaining the first and second weighing functions can be acquired in a preparatory phase prior to the acquisition process of the projection image or in a post-processing phase after the acquisition process of the projection image. The acquisition of the first and second gain images occurs manually or automatically.

According to an exemplary embodiment of the present invention the correction method allows selection of a degree at which the residual error of the scatter estimation is to be reduced. The correction unit operates in three different modes according to the selected degree of error reduction.

Corresponding to the three modes, the scatter correction method according to the present invention allows reducing the residual error of the estimated scattered distribution caused either solely by the ASG or solely by the non-uniformity of the source intensity, or by both. In the first mode the first weighing image is based on the first gain image. The second weighing image is taken to be proportional to the reciprocal of the first weighting image.

In the second mode, if the user wishes to account solely for the ASG shading, the first weighing image is set to unity and the second weighing image is based on the reciprocal of a pixel- wise ratio of the second gain image and the first gain image.

Finally, in the third mode, the first weighing image is derived from the first gain image and the second weighing image is taken to be proportional to the product of the reciprocal of the first weighting image and an image based on the reciprocal of the pixel- wise ratio of the second gain image and the first gain image.

The image format of the first and second gain image as well as the acquired projection image is immaterial for the purpose of the present invention as long as the format allows addressing each of the pixels to retrieve a corresponding pixel value. According to another exemplary embodiment of the present invention there is provided an imaging system comprising the above-mentioned correction unit.

According to yet another exemplary embodiment of the present invention there is provided a program element suitable for execution of the correction method according to the present invention. The present invention also provides a computer-readable medium for storing the inventive program element. BRIEF DESCRIPTION OF THE DRAWINGS

Aspects of a method, a device and a system in accordance with the present invention will be described in detail hereinafter with reference to the accompanying drawings therein

Fig. 1 shows a block diagram of an X-ray imaging system according to the present invention, Fig. 2 shows a flowchart of a scatter correction method according to the present invention.

DETAILED DESCRIPTION OF ASPECTS OF THE INVENTION The representations in the drawings are schematic and not to scale.

Fig. 1 shows basic components of an X-ray imaging system according to an exemplary embodiment of the present invention. The X-ray imaging system comprises an X-ray tube 1. The X-ray tube provides a source of X-ray radiation from which rays are emitted. The X-ray imaging system allows placing an object 5 between the X-ray tube 1 and a detector 12. Detector 12 has an entrance screen 11 and an anti-scatter-grid ASG arranged between the object 5 and the entrance screen 11. The anti-scatter-grid ASG is removably arranged on the detector 12.

The rays emitted from the X-ray tube 1 interact with the object 5 of which object 5 a projection image is to be acquired. The rays, after interaction with the object 5, and after having traversed the ASG are incident on the entrance screen 11. The incident rays are detected by the detector 12 and are transformed into light signals. The light signals upon exiting detector 12 via an exit screen 7 are acquired by an acquisition unit 13 such as a charged coupled device (CCD). The acquisition unit 13 acquires the light signals as a pattern of light intensities. The pattern of light intensities is passed on to an analog-digital-converter ADC. Dependent on the intensities the analog-digital-converter ADC outputs a pattern of pixel values arranged in an array, the pixel values giving a numerical indication of the intensities. The pixel values in the array are arranged in rows and columns. The pixel values correspond to points on the entrance screen 11 on which corresponding rays were incident at an intensity indicated by the pixel value. The converted array of pixel values is then stored on the storage unit 16 as a digital image file. The digital image file is the acquired projection image of the object 5. Alternatively or in addition to storing, the projection image is passed direct to a further post-processing unit 17 for further processing such as filtering. The projection image is then retrieved by the correction unit 19 to correct scatter artefacts in the acquired digital projection image. The correction unit can be arranged as a suitably programmed microchip such as an FPGA (field-programmable-gate-array) or as a software module running on a computer in communication with the imaging system, the software module being programmed in any suitable programming language such as MATLAB®, C or C++. The scatter corrected projection image is then forwarded to a monitor 15.

The projection image can then be visually examined on the monitor 15 or post-processed to reconstruct volumetric images.

Some of the rays are subject to scattering on their passage from the X-ray tube 1 via the object 5 to the entrance screen 11. It is because of the scattered rays that the anti-scatter-grid ASG is put into place. The anti-scatter-grid ASG comprises a number of lamellae focused on the X-ray tube 1 for shielding the entrance screen 11 from the scattered rays and letting only the unscattered and undeflected primary radiation reach the entrance screen 11. The shielding, however, is not perfect and the projection image is therefore a result of a superposition of a scatter- free primary radiation and the pure scatter radiation. The scatter artefacts in the projection image are due to the pure scattered radiation. The presence and features of the scatter artefacts are determined by a scatter distribution of the scattered radiation, the scatter distribution being detected by the detector 12.

The scatter distribution is affected by a spatial non-uniformity of the intensity of X-ray radiation emitted from the X-ray tube 1. Such non-uniformity results, for example, from heel effects, geometrical vignetting or from impacts due to a potential beam shaper. The scatter distribution is also affected by shading effects due to non- ideally aligned lamellae of the anti-scatter-grid ASG, which occurs, e.g., when the used distance between X-ray source 1 and the anti-scatter-grid ASG does not accurately equal the distance the ASG had originally be designed for. The described spatial non- uniformity of the X-ray source 1 and the anti-scatter-grid-shading are inherent "imperfections" in the image acquisition process commonly encountered in image acquisition systems. The spatial non-uniformity of the X-ray source 1 and the anti- scatter-grid-shading each affect the scatter distribution in a different manner and degree than the scatter- free primary radiation. According to an exemplary embodiment of the present invention at least one of a first gain image and a second gain image is acquired from the imaging system. The acquisitions of the first and second gain images are to account for the different manner and degree in which the spatial non-uniformity of the X-ray source 1 and the anti-scatter-grid-shading affect the scatter distribution. The gain images are acquired as a result of direct exposure of the detector 12 to the rays without the object 5 being arranged between the X-ray tube 1 and the detector 12. The gain images can be acquired either in a preparatory phase prior to acquiring the projection image of the object 5. The gain images can also be acquired in a phase after completion of acquiring the projection image and after the object 5 has been removed from the imaging system. Other than not having the object 5 arranged between the X-ray tube 1 and the detector 12, the acquisition of the gain images proceeds exactly as outlined for the case of the projection image. The second gain image is acquired with the ASG arranged on the entrance screen 11. The ASG needs to be removed prior to exposure of the detector to the rays in order to acquire the first gain image. The first or, as the case may be, the second gain image is then optionally forwarded to the post-processor 17 for filtering and for generating a first and a second weighting image to be used in correcting the projection image. The first and second gain image are stored on the data storage unit 16.

The operation of the correction unit 19 will be now explained in more detail by means of the flow-chart in Fig. 2.

The correction unit 19 takes as its main input the acquired projection image. Although as mentioned earlier the projection image can be thought of as a superposition of scatter- free primary radiation and a contribution due to scattered radiation the projection image can still be considered as good enough an approximation for the intensities of the primary radiation as to be used as the basis for an initial scatter estimation. The pixel values of the projection image as an approximation of the primary radiation intensity are available as values P (k, 1). The values P (k, 1) can be retrieved in step 20 as the intensity values in the projection image addressable by the column (k) and row position (1). The pair (k,l) is called pixel coordinates. In steps 22 and 24 a first weighing image and a second weighing image are retrieved on the basis of the stored first gain image and optionally the second gain image. The Intensity values in the first and second weighing images are also addressable by pixel coordinates k, 1 and are thus available as arrays of numbers w_imer (k, 1) and w_outer (i, j), respectively.

The scatter correction method according to the present invention estimates an X-ray scatter distribution S (i, j) in the imaging system, the estimation being based on convolution methods or convolution-type methods. The convolution-type methods are an extension of the convolution-method in that the convolution-type methods use a family of kernels rather than a single kernel for a convolution step.

A method for generating a family of kernels suitable for scatter correction purposes is disclosed by the applicant's international patent application WO2007/ 148263. The kernel generating method teaches calculation for said kernels in such a way that for each pixel in the projection image an asymmetric scatter distribution for error compensation is calculated. The asymmetric scatter distributions represent X- ray scatterings originating along a ray from the X-ray source to a position corresponding to a pixel position. The method also teaches simulating the X-ray scatterings according to geometric models - for example an ellipsoidal - having water-like scattering characteristics, each kernel of the family of kernels being a function of parameters of the geometrical model used.

The estimation of the X-ray scatter distribution S (i, j) according to the present invention proceeds along the following formula (Fl):

S(U) = A ^■ w_outei (i, j) ^• £ w_imer (*, /) ^• P(k, /) ^• κ_t, (i - kj - /) k,l

(Fl) Incorporating the two weighting images w_imer and w_outer into the convolutional step according to formula (Fl) allows for accounting for the technical imperfections during the image acquisition process affecting the scatter distribution. More specifically, the incorporation of the two weighting images w_imer and w_outer allows accounting for the non-uniform intensity of the radiation emitted by the X-ray source. The two weighting images further allow accounting for the shading effects caused by the anti-scatter-grid ASG.

The estimation of the scatter distribution S (i, j) according to the above formula is implemented in the correction unit 19 as a nested double loop, looping over the pixel coordinates k, 1. In formula (Fl), 'A' denotes an area per pixel.

As can be seen the estimation according to formula (Fl) is based on the convolution-type method mentioned earlier. According to one aspect of the present invention the values of the families of kernels Ky are predetermined and stored in the storage unit 16 as table values associated with pairs of indices (k,l) indicating a family member of the family of kernels Ky . The values are retrieved during execution according to formula (Fl) of the estimation step.

In step 26, the values of the members Ky of the family of kernels are weighted with P(k,l) and with the first weighting image w_ιτmer(k, //The values of the Ky represent scatter contributions generated along the rays incident on points of the entrance screen 11 corresponding to the pixel coordinates (k,l). The values of the Ky are weighted with the detected intensities P (k, 1) of the rays corresponding to the pixel coordinates (k,l).

The values of the Ky are further weighted by the first weighting image I) to account for the non-uniform intensity of the rays as emitted by the X-ray source 1.

The summation over the pixel coordinates (k,l) represents the weighted accumulation of all the scatter contributions generated along all the rays corresponding to the pixel coordinates (k,l) having the detected intensities P (k, 1).

According to an exemplary embodiment of the present invention, the correction method according to the present invention allows selectively accounting for either the non-uniformity of the X-ray source intensity, or the shading effect of the anti- scatter grid ASG shading or for both.

Experiments conducted by the applicant indicate reductions in the residual error of the scatter estimation according to the three selections. The reductions achieved are summarized in the following table: account ng for reduction in non-uniform non-uniform residual scatter ASG shading source intensity estimation error yes no 31% no yes 31% yes yes 50%

The selection affects how, according to the present invention, the two weighing images are to be based on the first and second gain images.

According to an exemplary embodiment of the present invention, the weighting images w_imer and w_outer are chosen to only compensate or account for the spatial non-uniformity of the X-ray intensity. In this mode, w_imer and w_outer are set as follows:

(F2)

^Wouter (ϊ J) =

-' source V' J)

In this mode, the step 28 of weighting with the second weighting function Wouter amounts to normalizing after weighting with the first weighting function w_mner in step 26.

Isource (k, 1) denotes the source intensity at the pixel coordinate (k, 1) derived from the first gain image. While processing formula (Fl) a routine can directly access the intensity value at the pixel positions (k, 1) and (i,j ) of the first and second weighting images w_irmer and w_outer Prior to retrieval in step 22 of the first and second weighting images, wwr and w_outer , the post-processor generates the weighting images according to formula (F2) on the basis of the source intensities I_SOUrce (k, 1) derived from the first gain image.

According to an exemplary embodiment of the present invention the weighting images wwr, w_outa are chosen so as to compensate only for the shading due to the anti-scatter-grid ASG. In this mode the weighting images w_irmer, w_outer are set as follows: w_imer (M) = I

C Λ ! (^F3)

⁷ASG (h j)

In formula (F3), I_ASG denotes intensity values obtained from a smoothed and/or filtered pixel- wise ratio image of the second gain image and the first image. Similarly as above in relation to formula (F2) the post-processor 17 in communication with the correction unit 19 generates the first and second weighting images w_imer, w_outer prior to retrieval in steps 22 and 24. The post-processor 17 accesses the first gain image and the second gain image and takes a ratio of the intensity values obtained from the second gain image and the first gain image. This ratio is then smoothed and/or filtered and stored at the pixel position i,j in the array I_ASG- In this second mode, the first weighing is set to unity and the step 28 of weighting with the second weighting function w_outer amounts to compensating for the shading effect of the anti-scatter-grid ASG.

According to an exemplary embodiment of the present invention the first and second weighting images, w_imer, w_outer, are chosen so as to compensate for both the non-uniformity of the X-ray intensity and the shading due to the ASG. In this mode the weighting images w_imer, w_outer are set according to formula: w_imer (k, /) = /_source (k, /) c _Λ 1 (F4)

Source (h j) ^' I ASG (h j)

In this mode, the step 28 of weighting with the second weighting function Wouter amounts to both, normalizing after weighting with the first weighting function Wmner and compensating for the shading effect of the anti-scatter-grid ASG.

In step 29 the projection image is corrected by looping through the pixels of the projection image and compensating for the estimated X-ray scatter distribution S(i,j), e.g. by subtraction, to obtain a corrected projection image.

The steps 26, 28 and 29 are iterated by feeding the corrected projection image as a next main input into the correction unit 19. The iteration loops through the sequence of corrected projection images obtained until a final corrected projection image is obtained which is deemed to have the contributions from scattered radiation removed at a sufficient level.

According to an exemplary embodiment of the present invention the image intensities IASG and I_SOUrce are calculated on the basis of models reflecting the parameters used in the imaging setup in the imaging system. More particularly, the intensity I_SOUrce is derived from a mathematical model of the emission characteristics of the X-ray tube 1. The Mathematical model would take into account the anode material, the anode angle and the applied tube voltage, etc. The Intensity I_ASG can be approximated by a geometrical model of the I_ASG used taking into account for example an aspect ratio and composition of the lamellae of the anti-scatter-grid ASG and the cover plate thickness, etc.

It should be noted that the term "comprising" does not exclude other elements or steps and the indefinite article "a" or "an" does not exclude the plural. Also elements described in association with different embodiments may be combined. It should also be noted that reference signs in the claims shall not be construed as limiting the scope of the claims.

Claims

CLAIMS:

1. An imaging system for acquiring a digital projection image of a physical object (5), the system comprising: an X-ray radiation source (1); a detector (12) for detecting rays of radiation originating from the radiation source (1); an acquisition unit (13) for acquiring the projection image; a data storage (16) for storing a first weighting image and a second weighting image; a correction unit (19) for X-ray scatter correction of the projection image, the correction unit (19) being configured to perform: based on the acquired projection image, estimating an X-ray scatter distribution, wherein a scatter contribution of at least one of the source rays is weighted with the first weighting image; weighting the estimated scatter distribution with the second weighting image; and correcting the projection image by using the weighted estimated scatter distribution.

2. The imaging system according to claim 1, wherein estimating the X-ray scatter distribution is based on a weighted accumulation of the scatter contribution generated along the at least one of the source rays.

3. The imaging system according to claims 1 or 2, wherein the weighting with the first weighting image accounts for a spatial variation of a source radiation intensity.

4. The imaging system according to claim 3 wherein the weighting with the second weighting image accounts for at least one of a normalization with respect to the spatial variation of the source radiation intensity and a shading effect caused by an anti- scatter grid (ASG).

5. The imaging system according to claim 4, wherein the first weighting image and the second weighting image are based on gain images acquired from the imaging system.

6. The imaging system according to claim 4, wherein the first weighting image is based on a first gain image acquired from the imaging system without the anti- scatter-grid arranged therein and wherein the second weighting image is based on a combination of the first weighting image and a pixel- wise ratio image of a second gain image acquired from the imaging system with having the anti-scatter grid (ASG) arranged therein and the first gain image.

7. A correction unit (19) for X-ray scatter correction of an acquired projection image, the acquisition of the projection image being based on detected rays of a radiation source (1), the correction unit (19) being configured to perform: based on the acquired projection image, estimating an X-ray scatter distribution, wherein a scatter contribution of at least one of the source rays is weighted with the first weighting image; weighting the estimated scatter distribution with the second weighting image; and correcting the projection image by using the weighted estimated scatter distribution.

8. A scatter correction method for correcting a digital projection image acquired from an imaging system, the acquisition of the projection image being based on detected rays of an X-ray radiation source (1) the method comprising: based on the acquired projection image, estimating an X-ray scatter distribution, wherein a scatter contribution of at least one of the source rays is weighted (26) with the first weighting image; weighting (28) the estimated scatter distribution with the second weighting image; and correcting (29) the projection image by using the weighted estimated scatter distribution.

9. The method according to claim 8, wherein estimating the X-ray scatter distribution is based on a weighted accumulation of the scatter contribution generated along the at least one of the source rays.

10. The method according to claim 8 or 9, wherein the weighting with the first weighting image accounts for a spatial variation of a source radiation intensity.

11. The method according to claim 10 wherein the weighting with the second weighting image accounts for at least one of a normalization with respect to the spatial variation of the source radiation intensity and a shading effect caused by an anti-scatter grid (ASG).

12. The method according to claim 11, wherein the first weighting image and the second weighting image are based on gain images acquired from the imaging system.

13. The method according to claim 11, wherein the first weighting image is based on a first gain image acquired from the imaging system without the anti-scatter- grid arranged therein and wherein the second weighting image is based on a combination of the first weighting image and a pixel- wise ratio image of a second gain image acquired from the imaging system with having the anti-scatter grid arranged therein and the first gain image.

14. A computer readable medium having stored thereon computer-executable instructions enabling a computer system to carry out a method for correcting a digital image, the method comprising: based on the acquired projection image, estimating an X-ray scatter distribution, wherein a scatter contribution of at least one of the source rays is weighted with the first weighting image; weighting the estimated scatter distribution with the second weighting image; and correcting the projection image by using the weighted estimated scatter distribution.

15. A program element configured and arranged to control when executed on a computer a method of scatter correction of a digital image, the method comprising: based on the acquired projection image, estimating an X-ray scatter distribution, wherein a scatter contribution of at least one of the source rays is weighted with the first weighting image; weighting the estimated scatter distribution with the second weighting image; and correcting the projection image by using the weighted estimated scatter distribution.