Method for correcting beam hardening projection, method for determining position of motif, related device and equipment
Through sector projection rearrangement and smooth convolution filtering technology, the beam hardening problem in X-ray computed tomography is identified and corrected, image processing efficiency is improved, and the inefficiency problem in the prior art is solved.
Patent Information
- Application Number
- CN202310555264.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-17
- Publication Date
- 2024-11-22
Smart Images

Figure CN119006622A_ABST
Abstract
Description
[0001] Nouns and pronouns relating to persons in this patent application are not limited to a specific gender. Technical Field
[0002] The present invention relates to computed tomography and in particular to image processing thereof. Background Art
[0003] The X-ray beam (referred to as beam) produced by the X-ray tube covers a wide spectrum. Since low-energy X-rays have a relatively high attenuation coefficient and are easily absorbed, high-energy X-rays become increasingly dominant, a phenomenon called beam hardening. Because of the beam hardening effect, the attenuation coefficient along the X-ray path is not constant, which will produce cup artifacts, shadows or stripe artifacts on the reconstructed image.
[0004] Beam filters can be used to reduce beam hardening. For CT, beam filters alone cannot meet all clinical needs, so additional algorithmic correction is required. Polynomial correction algorithms based on water phantoms are used because water is a major component of the human body. For each system, polynomial factors need to be determined. Because most methods for obtaining these factors require iteration or image reconstruction, and generally speaking, the iteration process and image reconstruction are inefficient, the entire beam hardening correction process is time-consuming. Due to the time cost, it is difficult or impossible to complete the entire correction procedure on each system.
[0005] The industry expects a more efficient method. Summary of the invention
[0006] In view of this, the present invention proposes a method for correcting beam hardening projection, a method for determining a phantom position, related devices, a computer program, a computer-readable storage medium and a device.
[0007] According to a first aspect of the present invention, there is provided a method for correcting a beam hardening projection, comprising:
[0008] Obtain fan-shaped projections of multiple phantoms of different sizes at various angles;
[0009] Rearranging the fan-shaped projections into equidistant parallel projections P;
[0010] According to the parallel projection P, identifying the position of the phantom;
[0011] Calculating theoretical parallel projections of the phantom according to the size, position and material of the phantom;
[0012] The fan projection is corrected according to the rearranged parallel projection P and the theoretical parallel projection.
[0013] In one embodiment, the steps include: obtaining fan-shaped projections of a plurality of phantoms of different sizes at various angles at the rotation center and at locations offset from the rotation center.
[0014] In one embodiment, the steps include:
[0015] For each channel, interpolating the fan-shaped projection along the projection angle direction, thereby obtaining non-equidistant parallel projections;
[0016] In the channel direction, the non-equidistant parallel projections are interpolated to obtain equidistant parallel projections P.
[0017] In one embodiment, the steps include:
[0018] Performing smoothing convolution filtering on the equidistant parallel projections P;
[0019] Determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 ;
[0020] According to the channel position E 1 and E 2 , calculate the position of the center of the phantom.
[0021] In one embodiment, the steps include performing smoothing convolution filtering according to the following formula:
[0022] FP=iFFT(FFT(P)×H(ω))
[0023] Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
[0024] In one embodiment, H(ω) is the product of a ramp function and a windowing function.
[0025] In one embodiment, the steps further include:
[0026] In the frequency domain, the projection after the smoothed convolution filter is padded with zeros.
[0027] In one embodiment, the step includes taking the channel that makes the filtered projection of the local minimum as the channel position E corresponding to the edge of the phantom 1 and E 2 .
[0028] In one embodiment, the steps include calculating the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula:
[0029]
[0030]
[0031]
[0032]
[0033]
[0034] in:
[0035] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0036] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0037] F is the size of the focal point;
[0038] r is the radius of the phantom;
[0039] F d is the distance from the center of the focus to the detector.
[0040] In one embodiment, the steps further include:
[0041] Fit a trajectory to the position of the center of the phantom.
[0042] In one embodiment, the phantom is a PMMA water phantom, and the steps include:
[0043] The attenuation per unit length of the water model is calculated according to the following formula:
[0044] μl=μ 1 (l 1 -l 2 )+μ 2 l 2
[0045] in:
[0046] μ 1 , μ 2 are the attenuation coefficients of PMMA and water, respectively;
[0047] l 1 , l 2 are the path length of the currently projected X-ray through the water phantom and the path length through the water, respectively;
[0048] μ is the attenuation per unit length of the water model;
[0049] l is the path length through the water model;
[0050] Based on μ and l, calculate the theoretical parallel projection.
[0051] In one embodiment, the steps include correcting the fan projection according to the following formula:
[0052]
[0053] in:
[0054] P WBC is the corrected fan projection;
[0055] B is the attenuation of the wedge filter;
[0056] a i,j is the water jet beam hardening correction factor.
[0057] According to a second aspect of the present invention, there is provided a method for determining a position of a phantom, comprising:
[0058] Obtaining fan-shaped projections of the phantom at various angles;
[0059] Rearranging the fan-shaped projections into equidistant parallel projections P;
[0060] Performing smoothing convolution filtering on the equidistant parallel projections P;
[0061] Determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 ;
[0062] According to the channel position E 1 and E 2 , calculate the position of the center of the phantom.
[0063] In one embodiment, the steps include: obtaining fan-shaped projections of the phantom at various angles at the rotation center and at locations deviating from the rotation center.
[0064] In one embodiment, the steps include:
[0065] For each channel, interpolating the fan-shaped projection along the projection angle direction, thereby obtaining non-equidistant parallel projections;
[0066] In the channel direction, the non-equidistant parallel projections are interpolated to obtain equidistant parallel projections P.
[0067] In one embodiment, the steps include performing smoothing convolution filtering according to the following formula:
[0068] FP=iFFT(FFT(P)×H(ω))
[0069] Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
[0070] In one embodiment, H(ω) is the product of a ramp function and a windowing function.
[0071] In one embodiment, the method further comprises:
[0072] In the frequency domain, the projection after the smoothed convolution filter is padded with zeros.
[0073] In one embodiment, the step includes taking the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom 1 and E 2 .
[0074] In one embodiment, the steps include calculating the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula:
[0075]
[0076]
[0077]
[0078]
[0079]
[0080] in:
[0081] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0082] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0083] F is the size of the focal point;
[0084] r is the radius of the phantom;
[0085] F d is the distance from the center of the focus to the detector.
[0086] In one embodiment, the method further comprises: fitting a trajectory of the position of the center of the phantom.
[0087] According to a third aspect of the present invention, there is provided a method for determining a position of a phantom, comprising:
[0088] Obtaining equidistant parallel projections P of the phantom at various angles;
[0089] Performing smoothing convolution filtering on the equidistant parallel projections P;
[0090] Determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 ;
[0091] According to the channel position E 1 and E 2 , calculate the position of the center of the phantom.
[0092] In one embodiment, the steps include: obtaining fan-shaped projections of the phantom at various angles at the rotation center and at locations deviating from the rotation center.
[0093] In one embodiment, the steps include performing smoothing convolution filtering according to the following formula:
[0094] FP=iFFT(FFT(P)×H(ω))
[0095] Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
[0096] In one embodiment, H(ω) is the product of a ramp function and a windowing function.
[0097] In one embodiment, the method further comprises:
[0098] In the frequency domain, the projection after the smoothed convolution filter is padded with zeros.
[0099] In one embodiment, the step includes taking the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom 1 and E 2 .
[0100] In one embodiment, the steps include calculating the channel position C corresponding to the center of the phantom according to the following formula:
[0101] In one embodiment, the method further comprises: fitting a trajectory of the position of the center of the phantom.
[0102] According to a fourth aspect of the present invention, there is provided a computer program, which, when executed by a processor, can implement the steps of the method described above.
[0103] According to a fifth aspect of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, wherein the program can implement the steps of the method described above when executed by a processor.
[0104] According to a sixth aspect of the present invention, there is provided an apparatus for correcting beam hardening projection, comprising:
[0105] A projection acquisition unit, configured to acquire fan-shaped projections of a plurality of phantoms of different sizes at various angles;
[0106] a rearrangement unit configured to rearrange the fan-shaped projections into equidistant parallel projections P;
[0107] a position identification unit configured to identify the position of the phantom according to the parallel projection;
[0108] a theoretical parallel projection calculation unit, configured to calculate the theoretical parallel projection of the phantom according to the size, position and material of the phantom;
[0109] The fan projection correction unit is configured to correct the fan projection according to the rearranged parallel projection P and the theoretical parallel projection.
[0110] In one embodiment, the projection acquisition unit is configured to acquire fan-shaped projections of a plurality of phantoms of different sizes at various angles located at the rotation center and at locations offset from the rotation center.
[0111] In one embodiment, the rearrangement unit comprises:
[0112] The non-equidistant parallel projection acquisition unit is configured to interpolate the fan-shaped projection along the projection angle direction for each channel respectively, so as to obtain non-equidistant parallel projections;
[0113] The equidistant parallel projection acquisition unit is configured to interpolate the non-equidistant parallel projections in the channel direction, so as to obtain the equidistant parallel projections P.
[0114] In one embodiment, the location identification unit includes:
[0115] A filtering unit, configured to perform smoothing convolution filtering on the equidistant parallel projections P;
[0116] A channel position determination unit is configured to determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 ;
[0117] The phantom position calculation unit is configured to calculate the position of the channel according to the channel position E 1 and E2 , calculate the position of the center of the phantom.
[0118] In one embodiment, the filtering unit comprises performing smoothing convolution filtering according to the following formula:
[0119] FP=iFFT(FFT(P)×H(ω))
[0120] Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
[0121] In one embodiment, H(ω) is the product of a ramp function and a windowing function.
[0122] In one embodiment, the location identification unit further includes:
[0123] The zero-padding unit is configured to fill zeros on the projection after the smoothing convolution filtering in the frequency domain.
[0124] In one embodiment, the channel position determination unit includes taking the channel that makes the filtered projection locally minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 .
[0125] In one embodiment, the phantom position calculation unit is configured to calculate the fan angle β of the center of the phantom and the distance L between the center of the phantom and the focus center according to the following formula:
[0126]
[0127]
[0128]
[0129]
[0130]
[0131] in:
[0132] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0133] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0134] F is the size of the focal point;
[0135] r is the radius of the phantom;
[0136] F d is the distance from the center of the focus to the detector.
[0137] In one embodiment, the location identification unit further includes:
[0138] A fitting unit is configured to fit a trajectory of the position of the center of the phantom.
[0139] In one embodiment, the phantom is a PMMA water phantom, and the theoretical parallel projection calculation unit includes:
[0140] The attenuation calculation unit is configured to calculate the attenuation per unit length of the water model according to the following formula:
[0141] μl=μ 1 (l 1 -l 2 )+μ 2 l 2
[0142] in:
[0143] μ 1 , μ 2 are the attenuation coefficients of PMMA and water, respectively;
[0144] l 1 , l 2 are the path length of the currently projected X-ray through the water phantom and the path length through the water, respectively;
[0145] μ is the attenuation per unit length of the water model;
[0146] l is the path length through the water model;
[0147] The second calculation unit is configured to calculate the theoretical parallel projection according to μ and l.
[0148] In one embodiment, the fan projection correction unit is configured to correct the fan projection according to the following formula:
[0149]
[0150] in:
[0151] P WBC is the corrected fan projection;
[0152] B is the attenuation of the wedge filter;
[0153] a i,j is the water jet beam hardening correction factor.
[0154] According to a seventh aspect of the present invention, there is provided a device for determining a position of a phantom, comprising:
[0155] A projection acquisition unit, configured to acquire fan-shaped projections of the phantom at various angles;
[0156] a rearrangement unit configured to rearrange the fan-shaped projections into equidistant parallel projections P;
[0157] A filtering unit, configured to perform smoothing convolution filtering on the equidistant parallel projections P;
[0158] A channel position determination unit is configured to determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 ;
[0159] The phantom position calculation unit is configured to calculate the position of the channel according to the channel position E 1 and E 2 , calculate the position of the center of the phantom.
[0160] In one embodiment, the projection acquisition unit is configured to acquire fan-shaped projections of the phantom at various angles at and away from the rotation center.
[0161] In one embodiment, the rearrangement unit comprises:
[0162] A non-equidistant parallel projection acquisition unit is configured to interpolate the fan-shaped projection along the projection angle direction for each channel, thereby obtaining non-equidistant parallel projections;
[0163] The equidistant parallel projection acquisition unit is configured to interpolate the non-equidistant parallel projections in the channel direction, so as to obtain the equidistant parallel projections P.
[0164] In one embodiment, the filtering unit is configured to perform smoothing convolution filtering according to the following formula:
[0165] FP=iFFT(FFT(P)×H(ω))
[0166] Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
[0167] In one embodiment, H(ω) is the product of a ramp function and a windowing function.
[0168] In one embodiment, the device further comprises: a zero-padding unit configured to fill zeros on the projection after the smoothing convolution filtering in the frequency domain.
[0169] In one embodiment, the channel position determination unit is configured to take the channel that makes the filtered projection locally minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 .
[0170] In one embodiment, the phantom position calculation unit is configured to calculate the fan angle β of the center of the phantom and the distance L between the center of the phantom and the focus center according to the following formula:
[0171]
[0172]
[0173]
[0174]
[0175]
[0176] in:
[0177] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0178] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0179] F is the size of the focal point;
[0180] r is the radius of the phantom;
[0181] F d is the distance from the center of the focus to the detector.
[0182] In one embodiment, the apparatus further comprises: a fitting unit configured to fit a trajectory of the position of the center of the phantom.
[0183] According to an eighth aspect of the present invention, there is provided a device for determining a position of a phantom, comprising:
[0184] A projection acquisition unit, configured to acquire equidistant parallel projections P of the phantom at various angles;
[0185] A filtering unit, configured to perform smoothing convolution filtering on the equidistant parallel projections P;
[0186] A channel position determination unit is configured to determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection1 and E 2 ;
[0187] The phantom position calculation unit is configured to calculate the position of the channel according to the channel position E 1 and E 2 , calculate the position of the center of the phantom.
[0188] In one embodiment, the projection acquisition unit is configured to acquire equidistant parallel projections P of the phantom at various angles located at the rotation center and at locations offset from the rotation center.
[0189] In one embodiment, the filtering unit is configured to perform smoothing convolution filtering according to the following formula:
[0190] FP=iFFT(FFT(P)×H(ω))
[0191] Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
[0192] In one embodiment, H(ω) is the product of a ramp function and a windowing function.
[0193] In one embodiment, the device further comprises: a zero-padding unit configured to fill zeros on the projection after the smoothing convolution filtering in the frequency domain.
[0194] In one embodiment, the channel position determination unit is configured to take the channel that makes the filtered projection locally minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 .
[0195] In one embodiment, the phantom position calculation unit is configured to calculate the channel position C corresponding to the center of the phantom according to the following formula:
[0196] In one embodiment, the method further comprises: a fitting unit configured to fit a trajectory of the position of the center of the phantom.
[0197] According to a ninth aspect of the present invention, there is provided a computer tomography device, comprising the computer readable storage medium as described above or the device as described above.
[0198] Compared with the image-based beam hardening correction algorithm, the method for correcting beam hardening projection, the method for determining the position of the phantom, the related apparatus, computer program, computer-readable storage medium and device of the present invention do not require time-consuming image reconstruction, and the efficiency advantage is obvious. For the projection-based beam correction algorithm, the key point is to identify the position of the phantom. Through the rearrangement from fan to parallel, the size of the phantom projection at all projections is changed to a known parameter, and the remaining phantom center position can be obtained by the edge enhancement method (which is not affected by beam hardening), making it more efficient. BRIEF DESCRIPTION OF THE DRAWINGS
[0199] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings so that the above and other features and advantages of the present invention will be more apparent to those skilled in the art. In the accompanying drawings:
[0200] Figure 1 FIG. 4 is a schematic flow chart of a method for correcting beam hardening according to an embodiment of the present invention.
[0201] Figure 2 for Figure 1 A schematic flow chart of the steps of the method is shown.
[0202] Figure 3 for Figure 1 A schematic flow chart of another step of the method is shown.
[0203] Figure 4A Schematic diagram of the attenuation value of the projection before filtering according to an embodiment of the present invention.
[0204] Figure 4B Schematic diagram of attenuation values of the filtered projection according to an embodiment of the present invention.
[0205] Figure 5 Schematic diagram of the attenuation value of the projection before and after filtering according to an embodiment of the present invention.
[0206] Fig. 6A FIG. 4 is a schematic diagram of the geometric relationship of a phantom at a certain projection according to an embodiment of the present invention.
[0207] Figure 6B This is a schematic diagram of the geometric relationship of the model at a certain projection when the focus is a point.
[0208] Figure 7 Schematic diagram of the trajectory and fitting of the center of the phantom according to an embodiment of the present invention.
[0209] Figure 8 for Figure 1 A schematic flow chart of yet another step of the method is shown.
[0210] Fig. 9Schematic diagram of the attenuation of a cylindrical water model.
[0211] Fig. 10A Schematic diagram of the measured projection of a 20 cm water phantom and the attenuation values of the WBC projection according to an embodiment of the present invention.
[0212] Fig. 10B Schematic diagram of the measured projection of a 30 cm water phantom and the attenuation values of the WBC projection according to an embodiment of the present invention.
[0213] Fig.11A is a schematic diagram of a reconstructed image of a 20 cm water phantom according to an embodiment of the present invention.
[0214] Fig. 11B is a schematic diagram of a reconstructed image of a 30 cm water phantom according to an embodiment of the present invention.
[0215] Fig.12 is a schematic flow chart of a method for determining a position of a phantom according to an embodiment of the present invention.
[0216] Fig.13 for Fig.12 A schematic flow chart of the steps of the method is shown.
[0217] Fig.14 is a schematic flow chart of a method for determining a position of a phantom according to an embodiment of the present invention.
[0218] Fig.15 FIG. 4 is a schematic structural block diagram of an apparatus for correcting beam hardening according to an embodiment of the present invention.
[0219] Fig.16 for Fig.15 A schematic structural block diagram of the rearrangement unit of the device shown.
[0220] Fig.17 for Fig.15 A schematic structural block diagram of a position identification unit of the device shown.
[0221] Fig.18 for Fig.15 A schematic structural block diagram of the theoretical parallel projection calculation unit of the device shown.
[0222] Fig.19 is a schematic structural block diagram of an apparatus for determining a position of a phantom according to an embodiment of the present invention.
[0223] Fig. 20 for Fig.19 A schematic flow chart of the rearrangement unit of the device shown.
[0224] Fig.21is a schematic structural block diagram of an apparatus for determining a position of a phantom according to an embodiment of the present invention.
[0225] In the above drawings, the reference numerals used are as follows:
[0226] 100 Method for correcting beam hardening 414; 514 Equidistant parallel projection acquisition unit
[0227] 200 Method for determining the position of a model 416; 516; 616 Filter unit
[0228] 300 Method for determining the position of the model 417; 517; 617 Zero filling unit
[0229] 400 Device for correcting beam hardening 418; 518; 618 Channel position determination unit
[0230] 402; 502; 602 projection acquisition unit 420; 520; 620 phantom position calculation unit
[0231] 404; 504 rearrangement unit 422; 522; 622 fitting unit
[0232] 406 Position identification unit 424 Attenuation calculation unit
[0233] 408 Theoretical parallel projection calculation unit 426 Second calculation unit
[0234] 410 sector projection correction unit 500; 600 device for determining the position of the phantom
[0235] 412; 512 Non-equidistant parallel projection acquisition unit 500; 600 Device for determining the position of the phantom
[0236] S102, S104, S106, S108, S110, S112, S114,
[0237] S116, S117, S118, S120, S122, S124, S126;
[0238] S202, S204, S212, S214, S216, S217, S218, steps
[0239] S220, S222; S302, S316, S317, S318, S320,
[0240] S322 DETAILED DESCRIPTION
[0241] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail with reference to the following embodiments.
[0242] Figure 1 1 is a schematic flow chart of a method 100 for correcting beam hardening according to an embodiment of the present invention. The method 100 for correcting beam hardening projection comprises steps S102, S104, S106, S108, and S110.
[0243] In step S102, sector projections of a plurality of phantoms of different sizes at various angles are obtained. Preferably, the phantom is a water phantom, because the density of water is close to that of the human body. Preferably, the outer shell of the phantom should be as thin as possible, or close to the density of water. To simulate the situation of human body scanning, phantoms of different sizes are scanned at the rotation center (for example, of a CT machine) and at locations deviated from the rotation center. Phantoms of different sizes have different corresponding correction factors. Phantoms with a diameter of 10-35 cm can be selected to cover different human body sizes. Depending on the size of the phantom, the phantom may deviate from the rotation center by 50-75 mm.
[0244] In step S104, the fan-shaped projections are rearranged into parallel projections P with equal distances. Figure 2 for Figure 1 FIG. 1 is a schematic flow chart of step S104 of method 100. Figure 2 As shown, in this embodiment, step S104 includes:
[0245] Step S112, interpolating the fan-shaped projections along the projection angle direction for each channel respectively, so as to obtain non-equidistant parallel projections;
[0246] Step S114: interpolate the non-equidistant parallel projections in the channel direction to obtain equidistant parallel projections P.
[0247] In step S106, the position of the phantom is identified according to the parallel projection. The present invention proposes an edge enhancement method to search for the position of the phantom. Figure 3 for Figure 1 FIG. 1 is a schematic flow chart of step S106 of method 100. Figure 3 As shown, in this embodiment, step S106 includes:
[0248] Step S116, performing smoothing convolution filtering on the equidistant parallel projections P. Step S116 includes performing smoothing convolution filtering according to the following formula:
[0249] FP=iFFT(FFT(P)×H(ω))
[0250] Wherein, FFT() is Fourier transform, H(ω) is a smoothing convolution filter function, and iFFT() is an inverse Fourier transform. In this embodiment, H(ω) is the product of a ramp function |ω| and a windowing function.
[0251] Figure 4A Schematic diagram of the attenuation value of the projection before filtering according to an embodiment of the present invention. Figure 4B Schematic diagram of attenuation values of a projection after filtering according to an embodiment of the present invention. Figure 4A and Figure 4B In the diagram, the longitudinal direction is the reading direction and the transverse direction is the channel direction. Figure 4A and Figure 4B As shown, after filtering, the edges of the model are enhanced, while other parts are weakened.
[0252] Step S118, determining the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 In this embodiment, step S118 includes taking the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 . Figure 5 Schematic diagram of the attenuation value of the projection before and after filtering according to an embodiment of the present invention, in which the solid line represents the projection after filtering, the dotted line represents the projection before filtering, and the channel position E corresponding to the minimum value circled in the figure 1 and E 2 It is the position of the edge projection of the model.
[0253] Step S120, according to the channel position E 1 and E 2 , calculate the position of the center of the model. Fig. 6A FIG. 4 is a schematic diagram of the geometric relationship of a phantom at a certain projection according to an embodiment of the present invention. Figure 6B This is a schematic diagram of the geometric relationship of the model at a certain projection when the focus is a point. Fig. 6A and Figure 6B , step S120 may include calculating the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula:
[0254]
[0255]
[0256]
[0257]
[0258]
[0259] in:
[0260] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0261] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0262] F is the size of the focal point;
[0263] r is the radius of the phantom;
[0264] F d is the distance from the center of the focus to the detector.
[0265] For parallel projection, the position of the phantom can be obtained in steps S116 and S118. In parallel projection, the channel position C corresponding to the center of the phantom can be obtained by the following formula:
[0266] Since there is usually noise in the data (for example, the limitation of the position accuracy by the size of the detector unit), in order to improve the accuracy of the position of the phantom, in this embodiment, step S106 may further include:
[0267] Step S122, fitting the trajectory of the position of the center of the phantom. In all projections, the fitting is the position of the phantom relative to the rotating part of the CT, and the fitting point of the first projection is the position of the phantom relative to the stationary part of the CT. Figure 7 Schematic diagram of the trajectory and fitting of the center of a phantom according to an embodiment of the present invention. The phantom has a diameter of 20 cm and is offset by 75 mm in the Y direction.
[0268] In order to improve the sampling density of the airspace, in this embodiment, step S106 may further include:
[0269] Step S117, in the frequency domain, zero-filling the projection after the smoothing convolution filtering. Zero-filling can improve the accuracy of the search position.
[0270] Step S108, calculating the theoretical parallel projection of the phantom according to the size, position and material of the phantom. Figure 8 for Figure 1 The schematic flow chart of step S108 of the method 100 shown in FIG. Fig. 9 Schematic diagram of attenuation of a cylindrical water phantom. In this embodiment, the phantom is a PMMA water phantom, and step S108 includes:
[0271] Step S124, calculating the attenuation per unit length of the water model according to the following formula:
[0272] μl=μ 1 (l 1-l 2 )+μ 2 l 2
[0273] in:
[0274] μ 1 , μ 2 are the attenuation coefficients of PMMA and water, respectively;
[0275] l 1 , l 2 are the path length of the currently projected X-ray through the water phantom and the path length through the water, respectively;
[0276] μ is the attenuation per unit length of the water model;
[0277] l is the path length through the water model, and l 1 ;
[0278] Step S126, calculating the theoretical parallel projection according to μ and l.
[0279] Step S110, correcting the fan projection according to the rearranged parallel projection P and the theoretical parallel projection. In this embodiment, step S110 may include correcting the fan projection according to the following formula:
[0280]
[0281] in:
[0282] P WBC is the corrected fan projection;
[0283] B is the attenuation of the wedge filter;
[0284] a i,j is the water jet beam hardening correction factor.
[0285] i+j≤3 is because using higher polynomial coefficients does not significantly improve the calculation accuracy.
[0286] Let V i,j =P·(P+B) i ·B j , then the equation in matrix form can be written as:
[0287]
[0288] in: is the theoretical parallel projection, k is the sampling point, and N is the number of sampling points, including multi-mode and multi-position data. The least squares fitting is used to solve the above matrix equation to obtain all a i,j, which are polynomial factors in the beam hardening correction formula. Although these factors are calculated from parallel projections, they can also be directly used for fan projections. Fig. 10A is a schematic diagram of the attenuation value of the measured projection and WBC projection of a 20 cm water phantom according to an embodiment of the present invention, Fig. 10B is a schematic diagram of the measured projection of the 30 cm water phantom and the attenuation value of the WBC projection according to an embodiment of the present invention. Fig. 10A and Fig. 10B The X-rays corresponding to the center part of the image have a larger travel path in the phantom and the beam hardening effect is more obvious, so the difference between its measured projection and the WBC projection is also greater. Fig.11A is a schematic diagram of a reconstructed image of a 20 cm water phantom according to an embodiment of the present invention. Fig. 11B Schematic diagram of a reconstructed image of a 30 cm water phantom according to an embodiment of the present invention. Fig.11A and Fig. 11B As shown, the CT values (means) at the edge and the middle points are comparable, indicating that calibration has been done.
[0289] Fig.12 2 is a schematic flow chart of a method 200 for determining a position of a phantom according to an embodiment of the present invention. The method 200 for determining a position of a phantom includes steps S202, S204, S216, S218, and S220.
[0290] In step S202, fan-shaped projections of the phantom at various angles are obtained.
[0291] In step S204, the fan-shaped projections are rearranged into parallel projections P with equal distances. Fig.13 for Fig.12 FIG. 2 is a schematic flow chart of step S204 of method 200. Fig.13 As shown, in this embodiment, step S204 includes:
[0292] Step S212, interpolating the fan-shaped projections along the projection angle direction for each channel respectively, so as to obtain non-equidistant parallel projections;
[0293] Step S214: interpolate the non-equidistant parallel projections in the channel direction to obtain equidistant parallel projections P.
[0294] Step S216, performing smoothing convolution filtering on the equidistant parallel projections P. Step S216 includes performing smoothing convolution filtering according to the following formula:
[0295] FP=iFFT(FFT(P)×H(ω))
[0296] Wherein, FFT() is Fourier transform, H(ω) is a smoothing convolution filter function, and iFFT() is an inverse Fourier transform. In this embodiment, H(ω) is the product of a ramp function |ω| and a windowing function.
[0297] Figure 4A Schematic diagram of the attenuation value of the projection before filtering according to an embodiment of the present invention. Figure 4B Schematic diagram of attenuation values of the filtered projection according to an embodiment of the present invention. Figure 4A and Figure 4B In the diagram, the longitudinal direction is the reading direction and the transverse direction is the channel direction. Figure 4A and Figure 4B As shown, after filtering, the edges of the model are enhanced, while other parts are weakened.
[0298] Step S218, determining the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 In this embodiment, step S218 includes taking the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 . Figure 5 Schematic diagram of the attenuation value of the projection before and after filtering according to an embodiment of the present invention, in which the solid line represents the projection after filtering, the dotted line represents the projection before filtering, and the channel position E corresponding to the minimum value circled in the figure 1 and E 2 It is the position of the edge projection of the model.
[0299] Step S220, according to the channel position E 1 and E 2 , calculate the position of the center of the model. Fig. 6A FIG. 4 is a schematic diagram of the geometric relationship of a phantom at a certain projection according to an embodiment of the present invention. Figure 6B This is a schematic diagram of the geometric relationship of the model at a certain projection when the focus is a point. Fig. 6A and Figure 6B , step S220 may include calculating the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula:
[0300]
[0301]
[0302]
[0303]
[0304]
[0305] in:
[0306] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0307] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0308] F is the size of the focal point;
[0309] r is the radius of the phantom;
[0310] F d is the distance from the center of the focus to the detector.
[0311] Since there is usually noise in the data (for example, caused by mechanical vibration), in order to improve the accuracy of the phantom position, in this embodiment, the method 200 may further include:
[0312] Step S222, fitting the trajectory of the position of the center of the phantom. In all projections, the fitting is the position of the phantom relative to the rotating part of the CT, and the fitting point of the first projection is the position of the phantom relative to the static part of the CT. Figure 7 Schematic diagram of the trajectory and fitting of the center of a phantom according to an embodiment of the present invention. The phantom has a diameter of 20 cm and is offset by 75 mm in the Y direction.
[0313] In order to improve the sampling density of the airspace, in this embodiment, the method 200 may further include:
[0314] Step S217: in the frequency domain, zero padding is performed on the projection after the smoothing convolution filtering. Zero padding can improve the accuracy of the search position.
[0315] Fig.14 3 is a schematic flow chart of a method 300 for determining a position of a phantom according to an embodiment of the present invention. The method 300 for determining a position of a phantom includes step S302, step S316, step S318, and step S320.
[0316] Step S302: Obtain equidistant parallel projections P of the phantom at various angles.
[0317] Step S316, performing smoothing convolution filtering on the equidistant parallel projections P. Step S316 includes performing smoothing convolution filtering according to the following formula:
[0318] FP=iFFT(FFT(P)×H(ω))
[0319] Wherein, FFT() is Fourier transform, H(ω) is a smoothing convolution filter function, and iFFT() is an inverse Fourier transform. In this embodiment, H(ω) is the product of a ramp function |ω| and a windowing function.
[0320] Figure 4A Schematic diagram of the attenuation value of the projection before filtering according to an embodiment of the present invention. Figure 4B Schematic diagram of attenuation values of the filtered projection according to an embodiment of the present invention. Figure 4A and Figure 4B In the diagram, the longitudinal direction is the reading direction and the transverse direction is the channel direction. Figure 4A and Figure 4B As shown, after filtering, the edges of the model are enhanced, while other parts are weakened.
[0321] Step S318, determining the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection 1 and E 2 In this embodiment, step S318 includes taking the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 . Figure 5 Schematic diagram of the attenuation value of the projection before and after filtering according to an embodiment of the present invention, in which the solid line represents the projection after filtering, the dotted line represents the projection before filtering, and the channel position E corresponding to the minimum value circled in the figure 1 and E 2 It is the position of the edge projection of the model.
[0322] Step S320, according to the channel position E 1 and E 2 , calculate the position of the center of the phantom. Step S320 includes calculating the channel position C corresponding to the center of the phantom according to the following formula:
[0323] Since there is usually noise in the data (for example, the limitation of the position accuracy by the size of the detector unit), in order to improve the accuracy of the position of the phantom, in this embodiment, the method 300 may further include:
[0324] Step S322, fitting the trajectory of the center position of the phantom. In all projections, the fitting is the position of the phantom relative to the rotating part of the CT, and the fitting point of the first projection is the position of the phantom relative to the stationary part of the CT. Figure 7 Schematic diagram of the trajectory and fitting of the center of a phantom according to an embodiment of the present invention. The phantom has a diameter of 20 cm and is offset by 75 mm in the Y direction.
[0325] In order to improve the sampling density of the airspace, in this embodiment, the method 300 may further include:
[0326] Step S317: In the frequency domain, the projection after the smoothing convolution filtering is filled with zeros. Zero filling can improve the accuracy of the search position.
[0327] The present invention also provides a computer program, which can implement the steps of method 100, method 200 or method 300 when executed by a processor.
[0328] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which can implement the steps of method 100, method 200 or method 300 when executed by a processor.
[0329] Fig.15 4 is a schematic structural block diagram of an apparatus 400 for correcting beam hardening according to an embodiment of the present invention. The apparatus 400 for correcting beam hardening projection includes a projection acquisition unit 402, a rearrangement unit 404, a position identification unit 406, a theoretical parallel projection calculation unit 408, and a fan projection correction unit 410.
[0330] The projection acquisition unit 402 is configured to acquire fan-shaped projections of a plurality of phantoms of different sizes at various angles. To simulate the situation of human body scanning, phantoms of different sizes are scanned at the rotation center (e.g., of a CT machine) and at locations offset from the rotation center. Phantoms of different sizes have different corresponding correction factors. Phantoms with a diameter of 10-35 cm can be selected to cover different human body sizes. Depending on the size of the phantom, the phantom can be offset from the rotation center by 50-75 mm.
[0331] The rearrangement unit 404 is configured to rearrange the fan-shaped projections into parallel projections P with equal distances. Fig.16 for Fig.15 FIG. 4 is a schematic structural block diagram of the rearrangement unit 404 of the device 400. Fig.16 As shown, in this embodiment, the rearrangement unit 404 includes:
[0332] The non-equidistant parallel projection acquisition unit 412 is configured to interpolate the fan-shaped projections along the projection angle direction for each channel, thereby obtaining non-equidistant parallel projections;
[0333] The equidistant parallel projection acquisition unit 414 is configured to interpolate the non-equidistant parallel projections in the channel direction, so as to obtain the equidistant parallel projections P.
[0334] The position identification unit 406 identifies the position of the phantom according to the parallel projection. The present invention proposes an edge enhancement method to search for the position of the phantom. Fig.17 for Fig.15 FIG. 4 is a schematic structural block diagram of the position identification unit 406 of the device 400. Fig.17As shown, in this embodiment, the position identification unit 406 includes:
[0335] The filtering unit 416 is configured to perform smoothing convolution filtering on the equidistant parallel projections P. The filtering unit 416 may perform smoothing convolution filtering according to the following formula:
[0336] FP=iFFT(FFT(P)×H(ω))
[0337] Wherein, FFT() is Fourier transform, H(ω) is a smoothing convolution filter function, and iFFT() is an inverse Fourier transform. In this embodiment, H(ω) is the product of a ramp function |ω| and a windowing function.
[0338] Figure 4A Schematic diagram of the attenuation value of the projection before filtering according to an embodiment of the present invention. Figure 4B Schematic diagram of attenuation values of the filtered projection according to an embodiment of the present invention. Figure 4A and Figure 4B In the diagram, the longitudinal direction is the reading direction and the transverse direction is the channel direction. Figure 4A and Figure 4B As shown, after filtering, the edges of the model are enhanced, while other parts are weakened.
[0339] The channel position determination unit 418 is configured to determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection. 1 and E 2 In this embodiment, the channel position determination unit 418 may use the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 . Figure 5 Schematic diagram of the attenuation value of the projection before and after filtering according to an embodiment of the present invention, in which the solid line represents the projection after filtering, the dotted line represents the projection before filtering, and the channel position E corresponding to the minimum value circled in the figure 1 and E 2 It is the position of the edge projection of the model.
[0340] The phantom position calculation unit 420 is configured to calculate the position of the channel E according to the 1 and E 2 , calculate the position of the center of the model. Fig. 6A FIG. 4 is a schematic diagram of the geometric relationship of a phantom at a certain projection according to an embodiment of the present invention. Figure 6B This is a schematic diagram of the geometric relationship of the model at a certain projection when the focus is a point. Fig. 6A and Figure 6B , the phantom position calculation unit 420 can calculate the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula:
[0341]
[0342]
[0343]
[0344]
[0345]
[0346] in:
[0347] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0348] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0349] F is the size of the focal point;
[0350] r is the radius of the phantom;
[0351] F d is the distance from the center of the focus to the detector.
[0352] For parallel projection, the position of the phantom can be obtained through the filtering unit 416 and the channel position determination unit 418 .
[0353] In parallel projection, the channel position C corresponding to the center of the phantom can be obtained by the following formula:
[0354] Since there is usually noise in the data (for example, the limitation of the position accuracy by the size of the detector unit), in order to improve the accuracy of the position of the phantom, in this embodiment, the position identification unit 406 may further include:
[0355] The fitting unit 422 is configured to fit the trajectory of the position of the center of the phantom. In all projections, the fitting is the position of the phantom relative to the rotating part of the CT, and the fitting point of the first projection is the position of the phantom relative to the static part of the CT. Figure 7 Schematic diagram of the trajectory and fitting of the center of a phantom according to an embodiment of the present invention. The phantom has a diameter of 20 cm and is offset by 75 mm in the Y direction.
[0356] In order to improve the sampling density of the airspace, in this embodiment, the location identification unit 406 may further include:
[0357] The zero-padding unit 417 fills zeros in the projection after the smoothing convolution filtering in the frequency domain. Zero-padding can improve the accuracy of the search position.
[0358] The theoretical parallel projection calculation unit 408 is configured to calculate the theoretical parallel projections of the phantom according to the size, position and material of the phantom. Fig.18 for Fig.15 The schematic structural block diagram of the theoretical parallel projection calculation unit 408 of the device 400 is shown. Fig. 9 Schematic diagram of attenuation of a cylindrical water phantom. In this embodiment, the phantom is a PMMA water phantom, and the theoretical parallel projection calculation unit 408 includes:
[0359] The attenuation calculation unit 424 is configured to calculate the attenuation per unit length of the water model according to the following formula:
[0360] μl=μ 1 (l 1 -l 2 )+μ 2 l 2
[0361] in:
[0362] μ 1 , μ 2 are the attenuation coefficients of PMMA and water, respectively;
[0363] l 1 , l 2 are the path length of the currently projected X-ray through the water phantom and the path length through the water, respectively;
[0364] μ is the attenuation per unit length of the water model;
[0365] l is the path length through the water model;
[0366] The second calculation unit 426 is configured to calculate the theoretical parallel projection according to μ and l.
[0367] The fan projection correction unit 410 is configured to correct the fan projection according to the rearranged parallel projection P and the theoretical parallel projection. In this embodiment, the fan projection correction unit 410 may correct the fan projection according to the following formula:
[0368]
[0369] in:
[0370] P WBC is the corrected fan projection;
[0371] B is the attenuation of the wedge filter;
[0372] a i,jis the water jet beam hardening correction factor.
[0373] i+j≤3 is because using higher polynomial coefficients does not significantly improve the calculation accuracy.
[0374] Let V i,j =P·(P+B) i ·B j , then the equation in matrix form can be written as:
[0375]
[0376] in: is the theoretical parallel projection, k is the sampling point, and N is the number of sampling points, including multi-modal and multi-position data.
[0377] Solving the above matrix equation using least squares fitting can obtain all a i,j , which are polynomial factors in the beam hardening correction formula. Although these factors are calculated from parallel projections, they can also be directly used for fan projections. Fig. 10A is a schematic diagram of the attenuation value of the measured projection and WBC projection of a 20 cm water phantom according to an embodiment of the present invention, Fig. 10B is a schematic diagram of the measured projection of the 30 cm water phantom and the attenuation value of the WBC projection according to an embodiment of the present invention. Fig. 10A and Fig. 10B The X-rays corresponding to the center part of the image have a larger travel path in the phantom, so the difference between its measured projection and the WBC projection is also larger. Fig.11A is a schematic diagram of a reconstructed image of a 20 cm water phantom according to an embodiment of the present invention. Fig. 11B Schematic diagram of a reconstructed image of a 30 cm water phantom according to an embodiment of the present invention. Fig.11A and Fig. 11B As shown, the CT values (means) at the edge and the middle points are comparable, indicating that calibration has been done.
[0378] Fig.19 5 is a schematic structural block diagram of an apparatus 500 for determining a phantom position according to an embodiment of the present invention. The apparatus 500 for determining a phantom position comprises a projection acquisition unit 502, a rearrangement unit 504, a filtering unit 516, a channel position determination unit 518, and a phantom position calculation unit 520.
[0379] The projection acquisition unit 502 is configured to acquire fan-shaped projections of the phantom at various angles.
[0380] The rearrangement unit 504 is configured to rearrange the fan-shaped projections into parallel projections P with equal distances. Fig. 20 for Fig.19 FIG. 5 is a schematic flow chart of the rearrangement unit 504 of the apparatus 500 shown in FIG. Fig. 20 As shown, in this embodiment, the rearrangement unit 504 includes:
[0381] The non-equidistant parallel projection acquisition unit 512 is configured to interpolate the fan-shaped projections along the projection angle direction for each channel, thereby obtaining non-equidistant parallel projections;
[0382] The equidistant parallel projection acquisition unit 514 is configured to interpolate the non-equidistant parallel projections in the channel direction, so as to obtain the equidistant parallel projections P.
[0383] The filtering unit 516 is configured to perform smoothing convolution filtering on the equidistant parallel projections P. The filtering unit 516 may perform smoothing convolution filtering according to the following formula:
[0384] FP=iFFT(FFT(P)×H(ω))
[0385] Wherein, FFT() is Fourier transform, H(ω) is a smoothing convolution filter function, and iFFT() is an inverse Fourier transform. In this embodiment, H(ω) is the product of a ramp function |ω| and a windowing function.
[0386] Figure 4A Schematic diagram of the attenuation value of the projection before filtering according to an embodiment of the present invention. Figure 4B Schematic diagram of attenuation values of the filtered projection according to an embodiment of the present invention. Figure 4A and Figure 4B In the diagram, the longitudinal direction is the reading direction and the transverse direction is the channel direction. Figure 4A and Figure 4B As shown, after filtering, the edges of the model are enhanced, while other parts are weakened.
[0387] The channel position determination unit 518 is configured to determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection. 1 and E 2 In this embodiment, the channel position determination unit 518 may use the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 . Figure 5 Schematic diagram of the attenuation value of the projection before and after filtering according to an embodiment of the present invention, in which the solid line represents the projection after filtering, the dotted line represents the projection before filtering, and the channel position E corresponding to the minimum value circled in the figure 1 and E 2 It is the position of the edge projection of the model.
[0388] The phantom position calculation unit 520 is configured to calculate the position of the channel according to the channel position E 1 and E 2, calculate the position of the center of the model. Fig. 6A FIG. 4 is a schematic diagram of the geometric relationship of a phantom at a certain projection according to an embodiment of the present invention. Figure 6B This is a schematic diagram of the geometric relationship of the model at a certain projection when the focus is a point. Fig. 6A and Figure 6B , the phantom position calculation unit 520 may include calculating the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula:
[0389]
[0390]
[0391]
[0392]
[0393]
[0394] in:
[0395] are the channel positions E corresponding to the edges of the model 1 and E 2 The fan angle;
[0396] They are the channel positions E corresponding to the edge of the model when the focus is a point. 1 and E 2 The fan angle;
[0397] F is the size of the focal point;
[0398] r is the radius of the phantom;
[0399] F d is the distance from the center of the focus to the detector.
[0400] Since there is usually noise in the data (for example, the limitation of the position accuracy by the size of the detector unit), in order to improve the accuracy of the position of the phantom, in this embodiment, the device 500 for determining the position of the phantom may further include:
[0401] The fitting unit 522 is configured to fit the trajectory of the position of the center of the phantom. In all projections, the fitting is the position of the phantom relative to the rotating part of the CT, and the fitting point of the first projection is the position of the phantom relative to the static part of the CT. Figure 7 Schematic diagram of the trajectory and fitting of the center of a phantom according to an embodiment of the present invention. The phantom has a diameter of 20 cm and is offset by 75 mm in the Y direction.
[0402] In order to improve the sampling density of the airspace, in this embodiment, the device 500 for determining the position of the phantom may further include:
[0403] The zero-padding unit 517 is configured to fill zeros on the projection after the smoothing convolution filtering in the frequency domain. Zero-padding can improve the accuracy of the search position.
[0404] Fig.21 6 is a schematic flow chart of an apparatus 600 for determining a phantom position according to an embodiment of the present invention. The apparatus 600 for determining a phantom position includes a projection acquisition unit 602 , a filtering unit 616 , a channel position determination unit 618 , and a phantom position calculation unit 620 .
[0405] The projection acquisition unit 602 is configured to acquire equidistant parallel projections P of the phantom at various angles.
[0406] The filtering unit 616 is configured to perform smoothing convolution filtering on the equidistant parallel projections P. The filtering unit 616 may perform smoothing convolution filtering according to the following formula:
[0407] FP=iFFT(FFT(P)×H(ω))
[0408] Wherein, FFT() is Fourier transform, H(ω) is a smoothing convolution filter function, and iFFT() is an inverse Fourier transform. In this embodiment, H(ω) is the product of a ramp function |ω| and a windowing function.
[0409] Figure 4A Schematic diagram of the attenuation value of the projection before filtering according to an embodiment of the present invention. Figure 4B Schematic diagram of attenuation values of the filtered projection according to an embodiment of the present invention. Figure 4A and Figure 4B In the diagram, the longitudinal direction is the reading direction and the transverse direction is the channel direction. Figure 4A and Figure 4B As shown, after filtering, the edges of the model are enhanced, while other parts are weakened.
[0410] The channel position determination unit 618 is configured to determine the channel position E corresponding to the edge of the phantom in the filtered equidistant parallel projection. 1 and E 2 In this embodiment, the channel position determination unit 618 may use the channel that makes the filtered projection local minimum as the channel position E corresponding to the edge of the phantom. 1 and E 2 . Figure 5 Schematic diagram of the attenuation value of the projection before and after filtering according to an embodiment of the present invention, in which the solid line represents the projection after filtering, the dotted line represents the projection before filtering, and the channel position E corresponding to the minimum value circled in the figure 1 and E2 It is the position of the edge projection of the model.
[0411] The phantom position calculation unit 620 is configured to calculate the position of the channel according to the channel position E 1 and E 2 , calculate the position of the center of the phantom. The phantom position calculation unit 620 can calculate the channel position C corresponding to the center of the phantom according to the following formula:
[0412] Since there is usually noise in the data (for example, the limitation of the position accuracy by the size of the detector unit), in order to improve the accuracy of the position of the phantom, in this embodiment, the device 600 for determining the position of the phantom may further include:
[0413] The fitting unit 622 is configured to fit the trajectory of the position of the center of the phantom. In all projections, the fitting is the position of the phantom relative to the rotating part of the CT, and the fitting point of the first projection is the position of the phantom relative to the static part of the CT. Figure 7 Schematic diagram of the trajectory and fitting of the center of a phantom according to an embodiment of the present invention. The phantom has a diameter of 20 cm and is offset by 75 mm in the Y direction.
[0414] In order to improve the sampling density of the airspace, in this embodiment, the device 600 for determining the position of the phantom may further include:
[0415] The zero padding unit 617 is configured to pad zeros on the projection after the smoothing convolution filtering in the frequency domain. Zero padding can improve the accuracy of the search position.
[0416] The present invention also provides a computer tomography device, comprising one or more of the following: the computer-readable storage medium described above, a device 400 for correcting beam hardening projection, a device 500 for determining a phantom position, and a device 600 for determining a phantom position.
[0417] Compared with the image-based beam hardening correction algorithm, the method for correcting beam hardening projection, the method for determining the position of the phantom, the related device, the computer program, the computer-readable storage medium and the equipment of the present invention do not require time-consuming image reconstruction, and the efficiency advantage is obvious. For the projection-based beam correction algorithm, the key point is to identify the position of the phantom. Through the rearrangement from fan to parallel, the size of the phantom projection at all projections is changed to a known parameter, and the remaining phantom center position can be obtained by the edge enhancement method (which is not affected by beam hardening).
[0418] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method (100) for correcting beam hardening projections, comprising: Step S102, obtaining fan-shaped projections of a plurality of phantoms of different sizes at various angles; Step S104, rearrange the fan-shaped projections into equidistant parallel projections P; Step S106, identifying the position of the phantom according to the parallel projection P; Step S108, calculating the theoretical parallel projection of the phantom according to the size, position and material of the phantom; Step S110, correcting the fan-shaped projection according to the rearranged parallel projection P and the theoretical parallel projection.
2. The method according to claim 1, characterized in that: Step S102 includes: obtaining fan-shaped projections of a plurality of phantoms of different sizes at various angles located at the rotation center and at locations deviating from the rotation center.
3. The method according to claim 1, characterized in that: Step S104 includes: Step S112, interpolating the fan-shaped projections along the projection angle direction for each channel respectively, so as to obtain non-equidistant parallel projections; Step S114: interpolate the non-equidistant parallel projections in the channel direction to obtain equidistant parallel projections P.
4. The method according to claim 1, characterized in that: Step S106 includes: Step S116, performing smoothing convolution filtering on the equidistant parallel projections P; Step S118, determining channel positions E1 and E2 corresponding to the edges of the phantom in the filtered equidistant parallel projections; Step S120: Calculate the position of the center of the phantom according to the channel positions E1 and E2.
5. The method according to claim 4, characterized in that: Step S116 includes performing smoothing convolution filtering according to the following formula: FP=iFFT(FFT(P)×H(ω)) Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
6. The method according to claim 5, characterized in that: H(ω) is the product of the ramp function and the windowing function.
7. The method according to claim 4, characterized in that: Step S106 also includes: Step S117, in the frequency domain, padding the projection after the smoothing convolution filtering with zeros.
8. The method according to claim 4, characterized in that: Step S118 includes taking the channel that makes the filtered projection locally minimum as the channel positions E1 and E2 corresponding to the edges of the phantom.
9. The method according to claim 4, characterized in that: Step S120 includes calculating the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula: in: are the fan angles of the channel positions E1 and E2 corresponding to the edges of the phantom, respectively; They are the fan angles of the channel positions E1 and E2 corresponding to the edge of the phantom when the focus is a point; F is the size of the focal point; r is the radius of the phantom; F d is the distance from the center of the focus to the detector.
10. The method according to claim 4, characterized in that: Step S106 also includes: Step S122, fitting the trajectory of the position of the center of the phantom.
11. The method according to claim 1, characterized in that: The phantom is a PMMA water phantom, and step S108 includes: Step S124, calculating the attenuation per unit length of the water model according to the following formula: μl=μ1(l1-l2)+μ2l2 in: μ1 and μ2 are the attenuation coefficients of PMMA and water, respectively; l1 and l2 are the path lengths of the currently projected X-ray through the water model and through the water, respectively; μ is the attenuation per unit length of the water model; l is the path length through the water model; Step S126, calculating the theoretical parallel projection according to μ and l.
12. The method according to claim 1, characterized in that: Step S110 includes correcting the fan projection according to the following formula: in: P WBC is the corrected fan projection; B is the attenuation of the wedge filter; a i,j is the water jet beam hardening correction factor.
13. A method (200) for determining a position of a phantom, comprising: Step S202, obtaining the fan-shaped projections of the phantom at various angles; Step S204, rearrange the fan-shaped projections into equidistant parallel projections P; Step S216, performing smoothing convolution filtering on the equidistant parallel projections P; Step S218, determining channel positions E1 and E2 corresponding to the edges of the phantom in the filtered equidistant parallel projections; Step S220: Calculate the position of the center of the phantom according to the channel positions E1 and E2.
14. The method according to claim 13, characterized in that: Step S202 includes: obtaining fan-shaped projections of the phantom at various angles at the rotation center and at locations deviating from the rotation center.
15. The method according to claim 13, characterized in that: Step S204 includes: Step S212, interpolating the fan-shaped projections along the projection angle direction for each channel respectively, so as to obtain non-equidistant parallel projections; Step S214: interpolate the non-equidistant parallel projections in the channel direction to obtain equidistant parallel projections P.
16. The method according to claim 13, characterized in that: Step S216 includes performing smoothing convolution filtering according to the following formula: FP=iFFT(FFT(P)×H(ω)) Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
17. The method according to claim 16, characterized in that: H(ω) is the product of the ramp function and the windowing function.
18. The method of claim 13, wherein: Also includes: Step S217, in the frequency domain, padding the projection after the smoothing convolution filtering with zeros.
19. The method of claim 13, wherein: Step S218 includes taking the channel that makes the filtered projection locally minimum as the channel positions E1 and E2 corresponding to the edges of the phantom.
20. The method of claim 13, wherein: Step S220 includes calculating the fan angle β of the center of the phantom and the distance L between the center of the phantom and the center of the focus according to the following formula: in: are the fan angles of the channel positions E1 and E2 corresponding to the edges of the phantom, respectively; They are the fan angles of the channel positions E1 and E2 corresponding to the edge of the phantom when the focus is a point; F is the size of the focal point; r is the radius of the phantom; F d is the distance from the center of the focus to the detector.
21. The method of claim 13, wherein: Also includes: Step S222, fitting the trajectory of the position of the center of the phantom.
22. A method (300) for determining a position of a phantom, comprising: Step S302, obtaining equidistant parallel projections P of the phantom at various angles; Step S316, performing smoothing convolution filtering on the equidistant parallel projections P; Step S318, determining channel positions E1 and E2 corresponding to the edges of the phantom in the filtered equidistant parallel projections; Step S320: Calculate the position of the center of the phantom according to the channel positions E1 and E2.
23. The method of claim 22, wherein: Step S302 includes: obtaining fan-shaped projections of the phantom at various angles at the rotation center and at locations deviating from the rotation center.
24. The method of claim 22, wherein: Step S316 includes performing smoothing convolution filtering according to the following formula: FP=iFFT(FFT(P)×H(ω)) Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
25. The method of claim 24, wherein: H(ω) is the product of the ramp function and the windowing function.
26. The method of claim 22, wherein: Also includes: Step S317, in the frequency domain, padding the projection after the smoothing convolution filtering with zeros.
27. The method of claim 22, wherein: Step S318 includes taking the channel that makes the filtered projection locally minimum as the channel positions E1 and E2 corresponding to the edges of the phantom.
28. The method of claim 22, wherein: Step S320 includes calculating the channel position C corresponding to the center of the phantom according to the following formula:
29. The method of claim 22, wherein: Also includes: Step S322, fitting the trajectory of the position of the center of the phantom.
30. A computer program, which, when executed by a processor, can implement the steps of the method according to any one of claims 1 to 29.
31. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 29 can be implemented.
32. An apparatus (400) for correcting beam hardening projection, comprising: A projection acquisition unit (402) is configured to acquire fan-shaped projections of a plurality of phantoms of different sizes at various angles; A rearrangement unit (404), configured to rearrange the fan-shaped projections into equidistant parallel projections P; A position identification unit (406) is configured to identify the position of the phantom according to the parallel projection; A theoretical parallel projection calculation unit (408), configured to calculate the theoretical parallel projection of the phantom according to the size, position and material of the phantom; The fan projection correction unit (410) is configured to correct the fan projection according to the rearranged parallel projection P and the theoretical parallel projection.
33. The device according to claim 32, characterized in that The projection acquisition unit (402) is configured to acquire fan-shaped projections of a plurality of phantoms of different sizes at various angles located at the rotation center and at locations deviated from the rotation center.
34. The device according to claim 32, characterized in that The rearrangement unit (404) comprises: A non-equidistant parallel projection acquisition unit (412) is configured to interpolate the fan-shaped projections along the projection angle direction for each channel, thereby obtaining non-equidistant parallel projections; The equidistant parallel projection acquisition unit (414) is configured to interpolate the non-equidistant parallel projections in the channel direction, thereby obtaining equidistant parallel projections P.
35. The device according to claim 32, characterized in that The position identification unit (406) comprises: A filtering unit (416) configured to perform smoothing convolution filtering on the equidistant parallel projections P; A channel position determination unit (418) is configured to determine channel positions E1 and E2 corresponding to the edges of the phantom in the filtered equidistant parallel projections; The phantom position calculation unit (420) is configured to calculate the position of the center of the phantom according to the channel positions E1 and E2.
36. The device according to claim 35, characterized in that The filtering unit (416) includes performing smoothing convolution filtering according to the following formula: FP=iFFT(FFT(P)×H(ω)) Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
37. The device according to claim 36, characterized in that H(ω) is the product of the ramp function and the windowing function.
38. The device according to claim 35, characterized in that The position identification unit (406) further includes: The zero padding unit (417) is configured to pad zeros on the projection after the smoothing convolution filtering in the frequency domain.
39. The device according to claim 35, characterized in that The channel position determination unit (418) includes taking the channel that makes the filtered projection locally minimum as the channel positions E1 and E2 corresponding to the edges of the phantom.
40. The device of claim 35, wherein: The phantom position calculation unit (420) is configured to calculate the fan angle β of the center of the phantom and the distance L between the center of the phantom and the focus center according to the following formula: in: are the fan angles of the channel positions E1 and E2 corresponding to the edges of the phantom, respectively; They are the fan angles of the channel positions E1 and E2 corresponding to the edge of the phantom when the focus is a point; F is the size of the focal point; r is the radius of the phantom; F d is the distance from the center of the focus to the detector.
41. The device according to claim 35, characterized in that The position identification unit (406) further includes: A fitting unit (422) is configured to fit a trajectory of the position of the center of the phantom.
42. The device according to claim 32, characterized in that The phantom is a PMMA water phantom, and the theoretical parallel projection calculation unit (408) includes: The attenuation calculation unit (424) is configured to calculate the attenuation per unit length of the water model according to the following formula: μl=μ1(l1-l2)+μ2l2 in: μ1 and μ2 are the attenuation coefficients of PMMA and water, respectively; l1 and l2 are the path lengths of the currently projected X-ray through the water model and through the water, respectively; μ is the attenuation per unit length of the water model; l is the path length through the water model; The second calculation unit (426) is configured to calculate the theoretical parallel projection according to μ and l.
43. The device according to claim 32, characterized in that The fan projection correction unit (410) is configured to correct the fan projection according to the following formula: in: P WBC is the corrected fan projection; B is the attenuation of the wedge filter; a i,j is the water jet beam hardening correction factor.
44. A device (500) for determining a position of a phantom, comprising: A projection acquisition unit (502) is configured to acquire fan-shaped projections of the phantom at various angles; A rearrangement unit (504), configured to rearrange the fan-shaped projections into equidistant parallel projections P; A filtering unit (516) configured to perform smoothing convolution filtering on the equidistant parallel projections P; A channel position determination unit (518) is configured to determine channel positions E1 and E2 corresponding to the edges of the phantom in the filtered equidistant parallel projections; The phantom position calculation unit (520) is configured to calculate the position of the center of the phantom according to the channel positions E1 and E2.
45. The device according to claim 44, characterized in that The projection acquisition unit (502) is configured to acquire fan-shaped projections of the phantom at various angles at and away from the rotation center.
46. The device according to claim 44, characterized in that The rearrangement unit (504) comprises: A non-equidistant parallel projection acquisition unit (512) is configured to interpolate the fan-shaped projections along the projection angle direction for each channel, thereby obtaining non-equidistant parallel projections; The equidistant parallel projection acquisition unit (514) is configured to interpolate the non-equidistant parallel projections in the channel direction, thereby obtaining equidistant parallel projections P.
47. The device according to claim 44, characterized in that The filtering unit (516) is configured to perform smoothing convolution filtering according to the following formula: FP=iFFT(FFT(P)×H(ω)) Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
48. The device according to claim 47, characterized in that H(ω) is the product of the ramp function and the windowing function.
49. The device according to claim 46, characterized in that include: The zero padding unit (517) is configured to pad zeros on the projection after the smoothing convolution filtering in the frequency domain.
50. The device of claim 44, wherein: The channel position determination unit (518) is configured to use the channel that makes the filtered projection locally minimum as the channel positions E1 and E2 corresponding to the edges of the phantom.
51. The device of claim 44, wherein: The phantom position calculation unit (520) is configured to calculate the fan angle β of the center of the phantom and the distance L between the center of the phantom and the focal center according to the following formula: in: are the fan angles of the channel positions E1 and E2 corresponding to the edges of the phantom, respectively; They are the fan angles of the channel positions E1 and E2 corresponding to the edge of the phantom when the focus is a point; F is the size of the focal point; r is the radius of the phantom; F d is the distance from the center of the focus to the detector.
52. The device of claim 44, further comprising: include: A fitting unit (522) is configured to fit a trajectory of the position of the center of the phantom.
53. A device (600) for determining a position of a phantom, comprising: A projection acquisition unit (602) is configured to acquire equidistant parallel projections P of the phantom at various angles; A filtering unit (616) configured to perform smoothing convolution filtering on the equidistant parallel projections P; A channel position determination unit (618) is configured to determine channel positions E1 and E2 corresponding to the edges of the phantom in the filtered equidistant parallel projections; The phantom position calculation unit (620) is configured to calculate the position of the center of the phantom according to the channel positions E1 and E2.
54. The device according to claim 53, characterized in that The projection acquisition unit (602) is configured to acquire equidistant parallel projections P of the phantom at various angles at and away from the rotation center.
55. The device according to claim 53, characterized in that The filtering unit (616) is configured to perform smoothing convolution filtering according to the following formula: FP=iFFT(FFT(P)×H(ω)) Among them, FFT() is Fourier transform, H(ω) is the smoothing convolution filter function, and iFFT() is the inverse Fourier transform.
56. The device according to claim 55, characterized in that H(ω) is the product of the ramp function and the windowing function.
57. The device according to claim 53, characterized in that include: The zero padding unit (617) is configured to pad zeros on the projection after the smoothing convolution filtering in the frequency domain.
58. The device of claim 53, wherein: The channel position determination unit (618) is configured to use the channel that makes the filtered projection locally minimum as the channel positions E1 and E2 corresponding to the edges of the phantom.
59. The device of claim 53, wherein: The phantom position calculation unit (620) is configured to calculate the channel position C corresponding to the center of the phantom according to the following formula:
60. The method of claim 53, further comprising: include: A fitting unit (622) is configured to fit a trajectory of the position of the center of the phantom.
61. A computer tomography apparatus comprising the computer readable storage medium of claim 31 or the apparatus of any one of claims 32 to 60.
Citation Information
Cited By
Method, device, computer program and storage medium for evaluating relative beam quality of X-ray
CN119087494A
Method and computer program for generating channel spectrum response difference correction table and channel spacing correction table
CN119423807B
