CT hardening artifact correction method for single-parameter quadratic function transformation

Abstract

The invention relates to the technical field of CT (Computed Tomography) imaging, in particular to a CT hardening artifact correction method for single-parameter quadratic function transformation, which comprises the following steps of: acquiring X-ray original image data of each angle of a CT system, and recording a projection angle sequence number and a probe position sequence number; dividing the original image by an air scanning background value, and then taking a negative logarithm to calculate a multi-energy projection value; selecting a data range according to an imaging type, for two-dimensional CT imaging, using a projection value at each angle, for three-dimensional cone beam CT imaging, only taking a middle line projection at each angle, and substituting the projection values into the optimization model to calculate corresponding parameters of single-parameter quadratic function transformation; selecting a correction formula according to positive and negative parameters, calculating correction projection by adopting quadratic function transformation when the parameters are positive, and ensuring monotonicity by adopting a correction formula when the parameters are not positive; and performing CT reconstruction on the corrected projection value to obtain an image after elimination of the hardening artifacts. According to the method, the correction process is simplified through single-parameter quadratic transformation, image edge gray scale abnormity is effectively inhibited, and hardening artifacts are weakened.

Description

Technical Field

[0001] This invention relates to the field of CT imaging technology, and specifically to a method for correcting CT hardening artifacts using a single-parameter quadratic function transformation. Background Technology

[0002] In conventional CT systems, X-rays are continuous multi-energy spectrum rays composed of photons of different energies. When passing through an object, low-energy rays are attenuated more easily than high-energy rays. Therefore, as the thickness increases, the average energy of the X-ray beam increases, resulting in "beam hardening," which causes hardening artifacts in the reconstructed image. Common software correction methods include multi-energy (dual-energy) imaging correction, linearization methods, and statistical iterative reconstruction. Multi-energy (dual-energy) imaging correction decomposes projection data under multiple different energy spectra to obtain a base projection image, and then synthesizes an approximate single-energy image. Linearization methods correct nonlinear data through linear transformations and require estimation using specific phantom projection data or other prior information.

[0003] Chinese invention patent application CN117665900A discloses an X-ray image correction method and system. The X-ray image correction method includes: simulating the X-ray energy spectrum distribution with tube voltage energy E of an X-ray tube; dividing the X-ray energy spectrum distribution into several single-energy sub-energy spectra according to the energy spectrum division step size; simulating the photon flux of X-rays of each sub-energy spectrum passing through the detected object with different penetration lengths L; converting the photon flux of the sub-energy spectrum into a projection value P; performing polynomial fitting between the penetration length L and the weighted projection value P of each sub-energy spectrum; and using the fitting curve of the penetration length L and the projection value P to correct the actual multi-energy projection value, thereby obtaining the ideal single-energy projection value and realizing the hardening correction of the X-ray image.

[0004] Polynomial fitting correction methods have strong universality. Essentially, they use polynomial functions to fit the transformation function from multi-energy projection to mono-energy projection. The multi-energy projection value should increase with the mono-energy projection value (or penetration thickness), exhibiting monotonicity. However, using high-order polynomials may cause oscillations in certain intervals, disrupting the monotonic correspondence between mono-energy projection (or penetration thickness) and multi-energy projection. High-order polynomial functions cannot guarantee their monotonicity, which may lead to the fitting results not accurately reflecting the correspondence between multi-energy projection values ​​and mono-energy projection values ​​(or penetration thickness). At the same time, polynomial fitting correction methods usually rely on correction phantoms, increasing implementation complexity. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the background technology by proposing a CT hardening artifact correction method based on a single-parameter quadratic function transformation.

[0006] The technical solution of this invention: a CT hardening artifact correction method based on single-parameter quadratic function transformation, comprising the following specific implementation steps:

[0007] S1. Acquire raw X-ray image data from each angle of the CT system. Set the acquisition angle number j = 1, 2, ..., J, where J is the total number of angles. Set the detector element position number i = 1, 2, ..., M for each angle, where M is the number of projected pixels per angle.

[0008] S2. Divide the original image by the background value I0 and take the negative logarithm. Calculate the projection value p of the j-th angle and the i-th element. ij And obtain the multi-energy projection value p;

[0009]

[0010] Among them, I ij The original grayscale image acquired for the j-th angle and the i-th probe;

[0011] S3. Select the projection data range according to the imaging type and solve for the single correction parameter a1:

[0012] For two-dimensional CT imaging, use all projection values ​​p. ij Substitute into the optimization model:

[0013]

[0014] If it is a three-dimensional cone-beam CT imaging, only the middle row projection value of the three-dimensional cone-beam CT of each angle detector is taken and substituted into the above model;

[0015] S4. Select the projection correction formula based on the sign of parameter a1:

[0016] If a1 > 0, then directly substitute it into the quadratic transformation formula:

[0017] p m =p 2 +a1p;

[0018] If a1≤0, then the corrected formula is used (to avoid non-monotonicity):

[0019] p m =(p-a1) 2 +a1(p-a1);

[0020] Where, p m The corrected projection value;

[0021] S5. Adjust the corrected projection value p m Perform CT reconstruction to obtain images after hardening artifacts have been eliminated.

[0022] Preferably, in step S3, the middle row projection value of the three-dimensional cone-beam CT is: the projection data of all angles when the detector row index is fixed at the middle position.

[0023] Preferably, in step S2, the projection value p ij This forms the projection matrix, i.e., the multi-energy projection value p.

[0024] Preferably, the background value I0 is determined by the grayscale value of the object-free X-ray projection image obtained by air scanning.

[0025] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects:

[0026] This invention presents a CT hardening artifact correction method based on a single-parameter quadratic function transformation. The single-parameter quadratic function transformation model significantly simplifies the hardening artifact correction process, requiring only the optimization of a single parameter to efficiently process projection data. Furthermore, it adaptively selects between full projection and intermediate projection calculations for both 2D and 3D cone-beam CT imaging, ensuring broad applicability. The method innovatively switches the correction formula dynamically based on the positive or negative value of the parameter. When the parameter is positive, a quadratic transformation is used; when the parameter is negative, a correction formula is activated, strictly guaranteeing the monotonicity of the transformation and avoiding artifact correction distortion. Ultimately, this method achieves stable elimination of hardening artifacts, improves the uniformity and accuracy of reconstructed images, and consumes low computational resources, making it easy to integrate into existing CT systems. Attached Figure Description

[0027] Figure 1 A flowchart of a CT hardening artifact correction method using a single-parameter quadratic function transform;

[0028] Figure 2 To directly reconstruct the image;

[0029] Figure 3 To directly reconstruct the grayscale changes at the position of the red line in the image;

[0030] Figure 4 Reconstructed image after hardening artifact correction;

[0031] Figure 5 The grayscale value change at the red line position in the reconstructed image after hardening artifact correction. Detailed Implementation

[0032] Example 1, as Figure 1 As shown, the CT hardening artifact correction method based on a single-parameter quadratic function transformation proposed in this invention includes the following specific implementation steps:

[0033] S1. Acquire raw X-ray image data from various angles of the CT system, and record the angle sequence number and detector position number. That is, use the CT scanning device to acquire X-ray images I of the object under test at different angles, set the acquisition angle sequence number j = 1, 2, ..., J (J is the total number of angles), and the detector element position number i = 1, 2, ..., M (M is the number of single-angle projection pixels) at each angle.

[0034] S2. Divide the original image by the background value and take the negative logarithm to calculate the multi-energy projection value p. ij :

[0035]

[0036] Where, p ij Ij represents the projection value of the j-th angle and the i-th detector element; I0 is the background value, i.e., the grayscale value of the X-ray image when there is no object (air scan); Ij represents the projection value of the j-th angle and the i-th detector element. ij The original grayscale image acquired for the j-th angle and the i-th probe;

[0037] S3. Select the data range according to the imaging type -- 2D CT uses all projection values, while 3D cone-beam CT only uses the middle row projection values. Substitute these values ​​into the optimization model to solve for the single parameter. Specifically:

[0038] (1) For two-dimensional CT imaging: use the projection values ​​p of all angles and probes. ij Substitute into the optimization model:

[0039]

[0040] Where a1 is a single correction parameter, i.e., the quadratic function transformation coefficient, which is solved by an optimization model;

[0041] (2) For three-dimensional cone-beam CT imaging: only the projection value p of the detector's middle row at each angle is taken. ij (Fixed middle row index i), substitute into the above formula to calculate a1;

[0042] S4. Determine the sign of the parameter: If the parameter is non-positive, use a quadratic function transformation to calculate the corrected projection; if the parameter is positive, use a correction formula to calculate the corrected projection to ensure monotonicity, specifically:

[0043] If a1 > 0, then directly substitute it into the quadratic transformation formula:

[0044] p m =p 2 +a1p;

[0045] If a1≤0, then the corrected formula is used (to avoid non-monotonicity):

[0046] p m =(p-a1)2 +a1(p-a1);

[0047] Where, p m p represents the corrected projection value; p represents the multi-energy projection value.

[0048] S5. Adjust the corrected projection value p m Perform CT reconstruction to obtain images after hardening artifacts have been eliminated.

[0049] Example 2: The CT hardening artifact correction method based on a single-parameter quadratic function transformation proposed in this invention is verified in certain specific scenarios as follows:

[0050] Taking an aluminum sample as an example;

[0051] Directly reconstructed images and grayscale changes at the red line position under a conventional CT system, as shown below. Figure 2 and Figure 3 As shown;

[0052] The CT hardening artifact correction method based on a single-parameter quadratic function transform proposed in Example 1 is shown in the reconstructed image and the grayscale changes at the red line position as follows: Figure 4 and Figure 5 As shown;

[0053] Therefore: In the directly reconstructed image, hardening artifacts are obvious, and the gray values ​​of pixels at the image edges are significantly higher than those in the middle. After hardening artifact correction, the difference in image gray values ​​becomes smaller.

[0054] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for correcting CT hardening artifacts using a single-parameter quadratic function transform, characterized in that, The specific implementation steps include the following: S1. Acquire raw X-ray image data from each angle of the CT system, set the acquisition angle number j = 1, 2, ..., J, where J is the total number of angles, and the detector element position number i = 1, 2, ..., M, where M is the number of projected pixels per angle; S2. Divide the original image by the background value I0 and take the negative logarithm. Calculate the projection value p of the j-th angle and the i-th element. ij And obtain the multi-energy projection value p; Among them, I ij The original grayscale image acquired for the j-th angle and the i-th probe; S3. Select the projection data range according to the imaging type and solve for the single correction parameter a1: For two-dimensional CT imaging, use all projection values ​​p. ij Substitute into the optimization model: If it is a three-dimensional cone-beam CT imaging, only the middle row projection value of the three-dimensional cone-beam CT of each angle detector is taken and substituted into the above model; S4. Select the projection correction formula based on the sign of parameter a1: If a1 > 0, then directly substitute it into the quadratic transformation formula: p m =p 2 +a1p; If a1≤0, then the corrected formula is used (to avoid non-monotonicity): p m =(p-a1) 2 +a1(p-a1); Where, p m The corrected projection value; S5. Perform CT reconstruction on the corrected projection value pm to obtain an image after hardening artifact correction.

2. The CT hardening artifact correction method based on a single-parameter quadratic function transform according to claim 1, characterized in that, In step S3, the projection value of the middle row of the three-dimensional cone-beam CT is: the projection data of all angles when the detector row index is fixed at the middle position.

3. The CT hardening artifact correction method based on a single-parameter quadratic function transform according to claim 1, characterized in that, The projection value p in step S2 ij This forms the projection matrix, i.e., the multi-energy projection value p.

4. The CT hardening artifact correction method based on a single-parameter quadratic function transform according to claim 1, characterized in that, The background value I0 is determined by the grayscale value of the objectless X-ray projection image obtained by air scanning.

Images