Uniformity correction method for response of detector of CT (Computed Tomography) device
By eccentric placement, semi-scan reconstruction and fan beam forward projection of the water mode image of the CT system, combined with polynomial fitting, the detector response uniformity correction coefficient is generated, which solves the problem of artifacts in the water mode image after the CT system correction. , achieving better image clarity and accuracy.
Patent Information
- Application Number
- CN202411068828.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-06
- Publication Date
- 2024-11-15
AI Technical Summary
After the system correction of the existing CT system, various artifacts still exist in the water-mode image, especially the asymmetric artifacts are difficult to effectively remove, and the existing methods have limited processing effects in the image domain.
By half-scanning reconstruction, smoothing, fan beam projection, and using polynomial fit to generate the detector response uniformity correction coefficient, channel to channel correction to eliminate the projected data deviation.
Effective removal of various artifacts, especially asymmetric artifacts, improve the clarity and accuracy of the image, simplify the correction process, and do not increase hardware costs.
Smart Images

Figure CN118967858A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method for correcting a system of an X-ray computer tomography device, and more specifically, to a method for correcting deviations that still exist in a CT system that has undergone system correction. Background Art
[0002] Water phantom images are often used to test the system performance of CT. Both national and industry standards mention the use of water phantom images to verify the performance of CT systems. These performances include image artifacts, CT value accuracy, CT image noise, CT value uniformity, etc.
[0003] The CT system is not an ideal system. For the same input X-ray, the response of each channel in the CT detector array cannot be consistent, and each channel has a unique gain coefficient.
[0004] X-rays are multi-energy spectrum rays. In the process of X-rays passing through an object and irradiating onto a detector, they are subject to scattering and beam hardening, which results in non-ideal rays reaching the detector. The purpose of the system correction of the X-ray computed tomography system is to eliminate the influence of the various actual hardware and physical effects mentioned above, obtain an ideal projection, and reconstruct an artifact-free image. Often, after a series of system corrections, including but not limited to intensity normalization correction, air correction (gain correction), scattering correction, crosstalk correction, hardening correction, and nonlinear correction, artifacts still exist in the water model image. The reason for the existence of the above-mentioned artifacts is that after the above-mentioned system correction, the projection data still has deviations compared with the ideal water model projection data. The purpose of the present invention is to correct the projection data as much as possible in a Channel to Channel manner to eliminate the water model image projection.
[0005] There are other solutions that can eliminate the existing artifacts on the water model image to a certain extent, such as the ring artifact removal algorithm, the dark band artifact removal algorithm, etc. These methods are all processed in the image domain and have requirements on the morphology of the artifacts. Generally, they are good for removing ring artifacts, but not so good for removing asymmetric artifacts. Summary of the invention
[0006] In view of the shortcomings of the prior art, the problem to be solved by the present invention is to remove various image artifacts caused by incomplete correction and deviation in system projection when there are no special requirements on the shape of water model image artifacts such as ring shape.
[0007] The technical solution adopted by the present invention to achieve the above-mentioned purpose is: a method for uniformity correction of detector response of a CT device, comprising the following steps:
[0008] For the corrected eccentrically placed water phantom projection, the water phantom image ImgOri with artifacts is obtained by half-scan reconstruction, and the corrected image ImgCor is obtained by smoothing.
[0009] The water phantom image ImgOri with artifacts and the corrected image ImgCor are respectively subjected to fan-beam forward projection to generate a projection with artifacts PrjOri and a projection without artifacts PrjIdeal;
[0010] For the projection PrjOri with artifacts, the projection values greater than the threshold are summed and averaged in the angle direction to obtain the smooth projection PrjOri_ViewAvg with artifacts. For the projection values of PrjIdeal corresponding to all angles where the PrjOri projection value is greater than the threshold, the smooth projection PrjIdeal_ViewAvg without artifacts is obtained.
[0011] Polynomial fitting is performed based on the smooth projection with artifacts PrjOri_ViewAvg and the smooth projection without artifacts PrjIdeal_ViewAvg to obtain the corresponding relationship between the two;
[0012] The water phantom projection data with artifacts to be corrected for uniformity is obtained, and the corrected water phantom projection data without artifacts is obtained through a relationship.
[0013] The eccentric placement means that the geometric center of the water model does not coincide with the scanning center of the frame.
[0014] The eccentric placement specifically means that the water model is placed to the edge of the scanning field of view but not beyond the scanning field of view.
[0015] The corrected eccentrically placed water model projection is achieved by completing the following corrections, including but not limited to: intensity normalization correction, air correction, scattering correction, crosstalk correction, hardening correction, and nonlinear correction.
[0016] The water phantom image ImgOri with artifacts is obtained by half-scan reconstruction, using data within the scanning range corresponding to (180°+α), where α is the fan beam angle.
[0017] The fan beam orthographic projection adopts parameters consistent with the actual geometric parameters of the CT system, including fan beam angle, focal position, and detector position; the starting angle of the fan beam orthographic projection is consistent with the starting angle of the scanning data used for the reconstructed water model image ImgOri.
[0018] The polynomial fitting is specifically as follows:
[0019] Take PrjOri_ViewAvg as the independent variable and PrjIdeal_ViewAvg as the dependent variable to perform polynomial fitting;
[0020] PrjIdeal_ViewAvg(c,s)
[0021] =a n (c,s)·PrjOri_ViewAvg(c,s)^n+a n-1 (c,s)·PrjOri_ViewAvg(c,s)^(n-1)
[0022] +…+a 2 (c,s)·PrjOri_ViewAvg(c,s)^2+a 1 (c,s)·PrJOri_ViewAvg(c,s)
[0023] +a 0 (c,s)
[0024] Among them, c is the detector channel direction index, s is the detector layer direction index, and n is the dimension of the polynomial fitting coefficient;
[0025] Get the detector response uniformity correction coefficient a n (c,s),a n-1 (c,s),…a 0 (c,s), c∈1,2,…C; s∈1,2,…S; where C is the number of channels of fan beam forward projection and S is the number of detector layers; each channel of the detector generates a set of polynomial coefficients.
[0026] A uniformity correction system for detector response of a CT device, comprising:
[0027] The half-scan reconstruction module is used to obtain the water phantom image ImgOri with artifacts through half-scan reconstruction for the corrected eccentrically placed water phantom projection, and perform smoothing to obtain the corrected image ImgCor;
[0028] The fan beam orthographic projection module is used to perform fan beam orthographic projection on the water model image ImgOri with artifacts and the corrected image ImgCor, respectively, to generate a projection with artifacts PrjOri and a projection without artifacts PrjIdeal;
[0029] A smooth projection module is used to add and average the projection values greater than a threshold in the projection PrjOri with artifacts in the angle direction to obtain a smooth projection PrjOri_ViewAvg with artifacts, and to add and average the projection values of PrjIdeal corresponding to all angles whose PrjOri projection values are greater than the threshold to obtain a smooth projection PrjIdeal_ViewAvg without artifacts;
[0030] A polynomial fitting module is used to perform polynomial fitting based on the smooth projection with artifacts PrjOri_ViewAvg and the smooth projection without artifacts PrjIdeal_ViewAvg to obtain the corresponding relationship between the two;
[0031] The uniformity correction module is used to obtain the water model projection data with artifacts to be corrected for uniformity, and obtain the corrected water model projection data without artifacts through a relationship.
[0032] A uniformity correction device for detector response of a CT device comprises a memory and a processor; the memory is used to store a computer program; the processor is used to implement the uniformity correction method for detector response of a CT device when executing the computer program.
[0033] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the uniformity correction method for the detector response of a CT device is implemented.
[0034] The present invention has the following beneficial effects and advantages:
[0035] 1. Each channel of the detector generates a set of correction coefficients. Compared with the artifact correction method in the image domain, the correction data is closer to the front end and has a better correction effect for artifacts of different forms.
[0036] 2. The calibration process is simple, and the phantom used is a common water phantom. Generally, CT systems are equipped with water phantoms as standard, which does not increase hardware costs.
[0037] 3. The reconstruction algorithm in the present invention is a two-dimensional half-scan reconstruction algorithm, which has the advantage that the reconstructed image can fully reflect the artifacts caused by the difference in detector response. The orthographic projection data with the same angle as the reconstructed data and the same geometric parameters as the real CT geometric parameters can reflect the detector response deviation in the orthographic projection data, thereby improving the correction effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 It is the sine diagram of the eccentric water model projection;
[0039] Figure 2 This is the water phantom image without uniformity correction;
[0040] Figure 3 Water phantom image for uniformity correction;
[0041] Figure 4 A flowchart for using a uniformity correction table during reconstruction;
[0042] Figure 5 Flowchart for generating detector uniformity correction table. DETAILED DESCRIPTION
[0043] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.
[0044] The present invention is referred to as a method for correcting the uniformity of the detector response of a CT device. Water phantom CT images are usually used to verify the correction effect of a CT system. The original detector data obtained by the CT system scanning the water phantom is corrected for data on the detector channel through intensity normalization correction, air correction, scattering correction, crosstalk correction, hardening correction, nonlinear correction, etc., and after image reconstruction, some CT system water phantom images still have bright bands, dark bands, bright rings or dark rings caused by channel differences. The present invention uses eccentric water phantom data to correct the raw data output by the CT detector after a series of corrections, and then further corrects the residuals that still exist by forward projecting the reconstructed eccentric water phantom image with artifacts and the corrected image, and generates a channel to channel detector uniformity correction table. (The detector uniformity correction table is also referred to as the uniformity correction table in the following text, and the correction method in the present invention is referred to as the detector uniformity correction, or simply referred to as the uniformity correction) After the uniformity correction table is generated, the uniformity correction table is used to correct the data during the scanning and reconstruction process, which can significantly reduce image artifacts.
[0045] The present invention is applicable to the case where there is still a deviation in the detector projection data after system correction, and there are still artifacts in the reconstructed water model image. Further correction is made to the system-corrected data to reduce image artifacts by correcting the detector deviation, including:
[0046] 1.1 The water model needs to be placed eccentrically. The purpose of eccentrically placing the water model is to allow the projection of the water model to cover more channels, so as to achieve the purpose of correcting more channels.
[0047] 1.2 Using the corrected eccentrically placed water phantom projection, half-scan reconstruction was performed to obtain a water phantom image with artifacts, recorded as ImgOri, and the image with artifacts was corrected by averaging, smoothing, etc., and recorded as ImgCor.
[0048] 1.3 The back-projection reconstruction mentioned in 1.2 is a half-scan reconstruction algorithm, that is, the reconstruction uses only the data within the scanning range corresponding to (180°+α), where α is the fan beam angle in degrees. Using half-scanning instead of full-scanning is a significant feature of the present invention, and its purpose is to ensure that there are no conjugate rays when the image is back-projected and reconstructed, so that the artifacts in the image can be strongly correlated with the channel response of the detector, and the reconstruction points will not be averaged by the conjugate channels and affected by the conjugate channels.
[0049] 1.4 The water phantom image ImgOri with artifacts and the corrected water phantom image ImgCor in 1.2 are images with the water phantom shell removed. For ImgOri and ImgCor mentioned in 1.2, fan beam orthographic projection is performed respectively. The orthographic projection geometric parameters should be consistent with the actual geometric parameters of the system, including but not limited to the fan beam angle, focus position, detector position and other specific parameters; the starting angle of the fan beam orthographic projection is consistent with the starting angle of the scanning data used to reconstruct the image ImgOri; combined with the description in 1.3, the artifacts in ImgOri are related to the scanning range, starting angle, and channels that still have errors after correction of the data that generates ImgOri. Therefore, fan beam orthographic projection according to the above description can completely simulate the characteristics of fan beam projection of ImgOri with artifacts, and this projection is recorded as PrjOri. Similarly, the same fan beam orthographic projection is performed on ImgCor. At this time, since there is no artifact in ImgCor, the projection generated by fan beam orthographic projection is also ideal, and this projection is recorded as PrjIdeal.
[0050] 1.5 Set the threshold Threshold. For projection values greater than Threshold in PrjOri, use WindowWidth as the window width to add and average the angle direction. The result is recorded as PrjOri_ViewAvg. For projections of PrjIdeal at the same position (index) as the projections greater than Threshold in PrjOri, use WindowWidth as the window width to add and average the angle direction. The result is recorded as PrjIdeal_ViewAvg. Use PrjOri_ViewAvg as the independent variable and PrjIdeal_ViewAvg as the dependent variable to perform polynomial fitting to fit the coefficient matrix of coefficient uniformity calibration. Where c is the channel direction index and s is the layer direction index.
[0051] PrjIdeal_ViewAvg(c,s)=
[0052] a n (c,s)·PrjOri_ViewAvg(c,s)^n+a n-1 (c,s)·PrjOri_ViewAvg(c,s)^(n-1)+…+a 2 (c,s)·PrjOri_ViewAvg(c,s)^2+a 1 (c,s)·PrJOri_ViewAvg(c,s)+a 0 (c,s)(1)
[0053] The water phantom image reconstructed by half scanning with artifacts is corrected. The data acquisition process of the corrected water phantom image and the water phantom image containing artifacts is simulated, and the projection data corresponding to the angle range of 180°+α is obtained, which are recorded as PrjIdeal and PrjOri respectively. The angle of the orthographic projection must be consistent with the data angle used to reconstruct the half-scan water phantom image. The threshold Threshold is set. For the projection values in PrjOri greater than the Threshold, the WindowWidth is used as the window width, and the angle direction is summed and averaged. The result is recorded as PrjOri_ViewAvg. For the projection of PrjIdeal at the same position (index) of the projection greater than the Threshold in PrjOri, the WindowWidth is used as the window width, and the angle direction is summed and averaged. The result is recorded as PrjIdeal_ViewAvg. With PrjOri_ViewAvg as the independent variable and PrjIdeal_ViewAvg as the dependent variable, a polynomial fitting is performed to fit the coefficient matrix of coefficient uniformity calibration, as shown in formula (1). The dimension of the polynomial coefficient is generally 2-dimensional, that is, n is 2. The polynomial coefficients used to correct the deviation can be obtained. The above polynomial coefficients are applied to the water phantom projection data corrected by the system during the reconstruction process to obtain a water phantom image with reduced artifacts.
[0054] In the above technical solution, it is necessary to collect eccentric water model data. One of the purposes of collecting eccentric water model is to allow the water model projection to cover as many channels as possible, so that the deviations of as many channels as possible can be corrected. Figure 1 As shown, the greater the eccentricity, the larger the channel range covered by the water model projection.
[0055] One feasible way to place the water model eccentrically mentioned above is to set different eccentric distances until the eccentric distance is 0, that is, the water model is placed at the center of the rack. The smaller the eccentric distance of the water model, the data collected is used to correct the channel close to the center, and vice versa, the larger the eccentric distance of the water model, the data collected is used to correct the data of the channel far from the center.
[0056] Another feasible way to place the water model eccentrically mentioned above is to place the water model to the edge of the scanning field of view, but not beyond the scanning field of view, and scan multiple laps of eccentric water model data.
[0057] Make necessary corrections to the eccentric water film data, including but not limited to intensity normalization correction, air correction (gain correction), scattering correction, crosstalk correction, defocus correction, water hardening correction, and nonlinear correction. After the above corrections, use the half-scan reconstruction algorithm to take the angle range data of PI+fan angle, perform half-scan reconstruction, and record the starting angle of the reconstructed data. Offset the starting angle by a certain angle, such as 90°, and then reconstruct again using the data corresponding to the angle of PI+fan angle. Repeat this operation to reconstruct N groups of images, and the reconstruction starting angle of each group of images is StartAngle[n], n=1,2,…N. Note that the reconstruction algorithm used in this article is a 2D half-scan reconstruction algorithm, that is, each image is reconstructed by only one layer of original data, unlike 3D reconstruction, where one image uses multiple layers of projection data, thereby simplifying the correction process.
[0058] The reconstructed N groups of images are processed as follows. The water shell of the water model image is removed, the air part is set to zero, and the CT value is converted into an attenuation value. The conversion process is performed according to formula (2), as described above. At this time, the image is recorded as ImgOri, where HU is the image CT value, μ water is the attenuation coefficient of water corresponding to the tube voltage kV used to scan the water model, and μ is the actual attenuation value corresponding to HU. ImgOri is processed. The optional method is to use the mean value of the water part in ImgOri to fill the water part in the image, thereby obtaining the water model image ImgCor. Thus, N groups of ImgOri and ImgCor are obtained, and the number of images in each group of ImgOri depends on the number of layers S of the original number.
[0059]
[0060] Using the geometric parameters of the CT system, fan beam orthographic projection is performed on ImgOri and ImgCor respectively. One of the key points is that the geometric parameters of CT need to be very accurate, and the system geometry needs to be modeled; the second key point is that the starting angle and projection angle range of the orthographic projection are consistent with the starting angle and scanning data range of the scan data used to reconstruct the image, and the angular direction sampling rate is consistent with the sampling rate of all scans. The above two points can ensure that the orthographic projection data corresponds one to one with the projection data obtained by the scan. The projection data obtained by the fan beam orthographic projection of ImgOri is recorded as PrjOri, and the projection data obtained by the fan beam orthographic projection of ImgCor is recorded as PrjIdeal. With PrjOri_ViewAvg as the independent variable and PrjIdeal_ViewAvg as the dependent variable, a polynomial fitting is performed. According to formula (1), the detector response uniformity correction coefficient a can be obtained. n (c,s),a n-1 (c,s),…a 0(c,s), c∈1,2,…C,s∈1,2,…S, where C is the number of channels of fan beam orthographic projection, which is consistent with the actual number of detector channels in the CT system, and S is the number of detector layers. Applying the uniformity correction coefficient to the reconstruction process can reduce artifacts caused by inconsistent detector response.
[0061] Figure 2 This is the water phantom image without detector response uniformity correction. Figure 3 This is a water phantom image for detector response uniformity correction. The display window width and window position are set to 50HU and 0HU.
[0062] The present invention uses a two-dimensional back-projection algorithm to obtain an image that is consistent with the number of detector layers. The reconstruction algorithm is a half-scan reconstruction algorithm. The reconstructed half-scan image is processed to remove artifacts in the image. Then, the fan-beam forward projection algorithm is used to obtain the forward projection of the artifact image and the forward projection of the artifact-free image. Finally, the correction coefficient of each channel is fitted. By using the above correction coefficient in reconstruction, image artifacts can be reduced. The process of generating uniformity correction coefficients is shown in Figure 5 , the process of using uniformity correction coefficients in the reconstruction process is shown in Figure 4 .
[0063] The present invention comprises the following steps:
[0064] S1, scanned eccentric water model raw data.
[0065] The eccentric water phantom should not exceed the scanning field of view. Because in the following steps, the water phantom image will be fan-beam orthographically projected, and the complete water phantom image should be orthographically projected. If the water phantom image is incomplete, the water phantom orthographic projection in the subsequent steps cannot correspond to the water phantom scanning data, resulting in the orthographic projection data not being able to fully contain the deviation information that causes the artifact.
[0066] The water model should be placed eccentrically to ensure that the lengths of the water models passed by the symmetrical channels of the detector are different; and in the subsequent steps, the angle range of the data used for half-scan reconstruction is used for front projection to ensure that the sum of the lengths of the water models passed by all views of a pair of symmetrical channels about the center channel is also not equal.
[0067] S2. Make necessary corrections to the eccentric water model data obtained by scanning.
[0068] These corrections include, but are not limited to, intensity normalization correction, air correction (gain correction), scatter correction, crosstalk correction, defocus correction, water hardening correction, and nonlinear correction. It should be noted that CT value correction, as an algorithm for calibrating CT value, is not included in the necessary correction algorithm described in S2.
[0069] S3. Reconstruct the image using a two-dimensional half-scan reconstruction algorithm for the water model data obtained in S2 after necessary correction.
[0070] One method to improve the uniformity correction effect is to reconstruct N groups of half-scanned images, with the starting angles of different groups of images differing by a certain angle. The N groups of images and the generated N groups of uniformity correction coefficients are averaged to improve the accuracy of the uniformity correction coefficients and avoid noise interference as much as possible.
[0071] The number of reconstructed images of a set of half-scan reconstructions is the same as the number of detector layers S. In a two-dimensional reconstruction algorithm, each layer of data produces an image. This allows the artifacts of each layer of image to uniquely correspond to each layer of data.
[0072] S4. The half-scanned image with artifacts generated in the previous step is processed to retain the pure water part. The CT value of the image is converted into a linear attenuation coefficient using formula (2). The image is recorded as ImgOri. The pure water part is smoothed and recorded as ImgCor.
[0073] The center of the water model can be found by the centroid method. The radius of the water model is known, and the thickness of the water model shell is known. Therefore, the image of the pure water part in the water model can be obtained.
[0074] The water model image ImgOri can be converted from rectangular coordinates to polar coordinates ImgOriPolar, and the polar coordinate water model image is smoothed along the polar radius direction to obtain a smoothed polar coordinate water model image, ImgCorPolar, and ImgCorPolar is converted to rectangular coordinates to obtain a smoothed water model image, ImgCor. In addition to the above method, the water part in ImgOri can be replaced by the mean value of water in ImgOri to obtain ImgCor. The same operation is performed for N groups of reconstructed images.
[0075] S5. Perform fan-beam orthographic projection on ImgCor and ImgOri respectively.
[0076] The geometric parameters of the CT used in fan-beam orthographic projection need to be very accurate, and the system geometry needs to be modeled so that the orthographic projection can truly simulate the actual scanning system.
[0077] The starting angle and projection angle range of the orthographic projection are consistent with the starting angle and scanning data range of the scan data used to reconstruct the image. As shown above, N sets of reconstructions are made, and the orthographic projection needs to be made for N sets of ImgCor and ImgOri respectively.
[0078] S6. As described above, the projection data obtained by the orthographic projection of the ImgOri fan beam is recorded as PrjOri, and the projection data obtained by the orthographic projection of the ImgCor fan beam is recorded as PrjIdeal. Set the threshold Threshold, and for the projection values in PrjOri that are greater than the Threshold, use WindowWidth as the window width to add and average the angles, and the result is recorded as PrjOri_ViewAvg. For the projections of PrjIdeal at the same position (index) as the projections in PrjOri that are greater than the Threshold, use WindowWidth as the window width to add and average the angles, and the result is recorded as PrjIdeal_ViewAvg. By calculating the summed average, high-frequency noise in the data can be eliminated to a certain extent, which is beneficial to the accuracy of data fitting. With PrjOri_ViewAvg as the independent variable and PrjIdeal_ViewAvg as the dependent variable, perform polynomial fitting, and now the detector response uniformity correction coefficient a can be obtained according to formula (1): n (c,s),a n-1 (c,s),…a 0 (c,s), c∈1,2,…C,s∈1,2,…S, where C is the number of channels of fan beam orthographic projection, which is consistent with the actual number of detector channels in the CT system, and S is the number of detector layers.
[0079] For a set of orthographic projections of a water model image, a set of correction coefficients is generated. The orthographic projections of N sets of reconstructed images can be polynomially fitted in the above manner to generate correction coefficients, and the N sets of correction coefficients are averaged to obtain the final detector uniformity correction coefficients, which can further eliminate the errors caused by the fitting process.
[0080] In addition to the system calibration and the detector response uniformity calibration mentioned in the present invention, the system needs to be recalibrated for CT value to ensure the accuracy of the CT value.
[0081] The procedure for applying detector uniformity correction during the reconstruction process is shown in Figure 4 .
Claims
1. A method for uniformity correction of detector response of a CT device, It is characterized in that The following steps are involved: For the corrected eccentrically placed water phantom projection, the water phantom image ImgOri with artifacts is obtained by half-scan reconstruction, and the corrected image ImgCor is obtained by smoothing. The water phantom image ImgOri with artifacts and the corrected image ImgCor are respectively subjected to fan-beam forward projection to generate a projection with artifacts PrjOri and a projection without artifacts PrjIdeal; For the projection PrjOri with artifacts, the projection values greater than the threshold are summed and averaged in the angle direction to obtain the smooth projection PrjOri_ViewAvg with artifacts. For the projection values of PrjIdeal corresponding to all angles where the PrjOri projection value is greater than the threshold, the smooth projection PrjIdeal_ViewAvg without artifacts is obtained. Polynomial fitting is performed based on the smooth projection with artifacts PrjOri_ViewAvg and the smooth projection without artifacts PrjIdeal_ViewAvg to obtain the corresponding relationship between the two; The water phantom projection data with artifacts to be corrected for uniformity is obtained, and the corrected water phantom projection data without artifacts is obtained through a relationship.
2. A method for uniformity correction of detector response of a CT device according to claim 1, It is characterized in that The eccentric placement means that the geometric center of the water model does not coincide with the scanning center of the frame.
3. The method for uniformity correction of detector response of a CT device according to claim 1, It is characterized in that The eccentric placement specifically means that the water model is placed to the edge of the scanning field of view but not beyond the scanning field of view.
4. The method for uniformity correction of detector response of a CT device according to claim 1, It is characterized in that The corrected eccentrically placed water model projection is achieved by completing the following corrections, including but not limited to: intensity normalization correction, air correction, scattering correction, crosstalk correction, hardening correction, and nonlinear correction.
5. The method for uniformity correction of detector response of a CT device according to claim 1, It is characterized in that The water phantom image ImgOri with artifacts is obtained by half-scan reconstruction, using data within the scanning range corresponding to (180°+α), where α is the fan beam angle.
6. A method for uniformity correction of detector response of a CT device according to claim 1, It is characterized in that The fan beam orthographic projection adopts parameters consistent with the actual geometric parameters of the CT system, including fan beam angle, focal position, and detector position; the starting angle of the fan beam orthographic projection is consistent with the starting angle of the scanning data used for the reconstructed water model image ImgOri.
7. The method for uniformity correction of detector response of a CT device according to claim 1, It is characterized in that The polynomial fitting is specifically as follows: Take PrjOri_ViewAvg as the independent variable and PrjIdeal_ViewAvg as the dependent variable to perform polynomial fitting; PrjIdeal_ViewAvg(c,s)=a n (c,s)·PrjOri_ViewAvg(c,s)^n+a n-1 (c,s)·PrjOri_ViewAvg(c,s)^(n-1)+…+a 2 (c,s)·PrjOri_ViewAvg(c,s)^2+a 1 (c,s)·PrJOri_ViewAvg(c,s)+a 0 (c,s) Among them, c is the detector channel direction index, s is the detector layer direction index, and n is the dimension of the polynomial fitting coefficient; Get the detector response uniformity correction coefficient a n (c,s),a n-1 (c,s),…a 0 (c,s), c∈1,2,…C; s∈1,2,…S; where C is the number of channels of fan beam forward projection and S is the number of detector layers; each channel of the detector generates a set of polynomial coefficients.
8. A uniformity correction system for detector response of a CT device, It is characterized in that include: The half-scan reconstruction module is used to obtain the water phantom image ImgOri with artifacts through half-scan reconstruction for the corrected eccentrically placed water phantom projection, and perform smoothing to obtain the corrected image ImgCor; The fan beam orthographic projection module is used to perform fan beam orthographic projection on the water model image ImgOri with artifacts and the corrected image ImgCor, respectively, to generate a projection with artifacts PrjOri and a projection without artifacts PrjIdeal; A smooth projection module is used to add and average the projection values greater than a threshold in the projection PrjOri with artifacts in the angle direction to obtain a smooth projection PrjOri_ViewAvg with artifacts, and to add and average the projection values of PrjIdeal corresponding to all angles whose PrjOri projection values are greater than the threshold to obtain a smooth projection PrjIdeal_ViewAvg without artifacts; A polynomial fitting module is used to perform polynomial fitting based on the smooth projection with artifacts PrjOri_ViewAvg and the smooth projection without artifacts PrjIdeal_ViewAvg to obtain the corresponding relationship between the two; The uniformity correction module is used to obtain the water model projection data with artifacts to be corrected for uniformity, and obtain the corrected water model projection data without artifacts through a relationship.
9. A device for correcting the uniformity of a detector response of a CT device, It is characterized in that The invention comprises a memory and a processor; the memory is used to store a computer program; the processor is used to implement a uniformity correction method for detector response of a CT device as claimed in any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, It is characterized in that The storage medium stores a computer program, and when the computer program is executed by a processor, the method for uniformity correction of detector response of a CT device as claimed in any one of claims 1 to 7 is implemented.
