JP4416545B2 - X-ray CT system - Google Patents

Description

  The present invention relates to noise removal in an X-ray CT apparatus that analyzes the internal structure of a subject by detecting and reproducing the X-ray absorption distribution of the subject by irradiating the subject with X-rays.

In parts where bones such as shoulders and pelvis overlap, or where X-ray passes are long, X-ray absorption is greater than other parts, and the level of transmitted X-ray dose is very low. As a result, the S / N of projection data The ratio gets worse. If an image is reconstructed with projection data including data with such a poor S / N ratio, it causes artifacts in the image. FIG. 1 is a diagram in which an image is reconstructed using data 11 of a shoulder portion including noise. Conventionally, for example, Patent Document 1 proposes to obtain an image with reduced artifacts 12 by performing noise removal by smoothing a portion where the X-ray dose level is low, that is, a portion where the projection data level is high. .
JP 2000-5159 specification

  According to Patent Document 1, there is a possibility that an actual data change portion that is not noise may be erased by performing smoothing, resulting in a decrease in resolution.

  An object of the present invention is to provide an X-ray CT apparatus having a function of detecting and removing noise components that cause artifacts mixed in transmitted X-ray data without causing a decrease in resolution in the X-ray CT apparatus. That is.

  Another object of the present invention is to detect and remove at high speed noise components that cause artifacts mixed in transmitted X-ray data without causing a reduction in resolution in an X-ray CT apparatus.

  The principle of the present invention is to detect the presence or absence of noise from the fluctuation of the total value for each view angle of the projection data, perform smoothing to remove the noise, and make the total value of the projection data equal to the smoothing data of the view angle The noise is removed by correcting each projection data so that

  A detector composed of a large number of detector elements arranged in an arc shape for detecting the transmitted X-ray dose has two-dimensional different characteristics in the channel direction and view direction in the sinogram, but the sum value of the projection data in the channel direction is the same. If measured by a cross section, it is constant regardless of the view, and in a helical scan in which projection is performed by moving the table, the sum value of the projection data does not become constant but changes continuously. Therefore, the fluctuation of the Sum value is considered to be the influence of noise depending on the view angle. By smoothing up and down the entire Sum value curve plotted along the view angle so that the Sum value is constant or continuous, the data characteristics in the channel direction are maintained, so noise components can be removed without causing a decrease in resolution. .

  According to a first aspect of the present invention, in the X-ray CT apparatus, fluctuations in the total value (Sum value) for each view angle of data (raw data) obtained by converting measurement noise in the view direction from X-ray dose data to projection data. From this, the presence or absence of noise is estimated and detected, X-ray dose data before conversion is corrected, and projection data is obtained again, thereby removing noise without causing a reduction in resolution.

  According to a second aspect of the present invention, in an X-ray CT apparatus, X-ray dose from raw data is preliminarily applied to data that needs to be corrected for X-ray dose that is data before conversion based on projection data information. Time required for data correction for noise removal by using a table of raw data correction amount that assumes and reflects the steps of reverse conversion to X-ray, X-ray dose correction, and re-conversion from corrected X-ray dose to raw data And the signal processing time is increased.

  Noise is mixed in the X-ray dose before conversion to raw data. Further, since the conversion from the X-ray dose to the raw data includes a non-linear logarithmic function, it is desirable to perform noise correction on the first transmitted X-ray dose data. For this reason, in the first aspect of the present invention, the X-ray dose data before conversion is corrected and converted again to raw data. However, this takes time for conversion and processing. Therefore, in the second aspect of the present invention, assuming the steps of inverse conversion, correction, and reconversion in advance, the correction amount of the reflected raw data is tabulated and converted into a table. By using it, the processing time was increased.

  FIG. 2 is a diagram for explaining the outline of the X-ray CT apparatus data collection and the sinogram generated from the collected data. While rotating the scanner, X-rays are generated from the X-ray source 21, and the X-ray dose transmitted through the subject 22 is measured by the detector 23. The measured data is subjected to processing such as offset correction 31, logarithmic conversion 32, reference correction 33 and the like in the image processing apparatus 50 as shown in the flowchart of FIG. 3, and as a result, raw data 24 is created. The image processing apparatus 50 includes, for example, a CPU, POM, RAM, and the like.

  The offset correction 31 is a process of subtracting the data amount measured when the X-ray is not irradiated from the measurement data, and the result corresponds to the X-ray dose data.

In general, attenuation of X-rays shows an exponential change with respect to the distance that X-rays pass through a substance, and thus a logarithmic function is used for calculation to obtain raw data from X-ray dose. Further, the reference correction is performed by adjusting a constant determined for each view angle so that the raw data becomes a certain level or more regardless of the irradiation dose. These can be summarized in [Equation 1].
It becomes.

  FIG. 4 shows a value (Sum value) obtained by summing the raw data at each view angle at the time of measurement in the channel direction. The Sum value of the projection data is constant regardless of the view angle as long as it is measured on the same cross section. In the spiral scan in which the projection is performed by moving the table, the Sum value of the projection data does not become constant but continuously changes (no noise 41). Therefore, the Sum value does not become discontinuous unless noise is mixed.

However, when noise is mixed, the Sum value shows a discontinuous change such as 42 with noise.
In the present invention, first, a change for each view angle of the Sum value is plotted and smoothing as shown in [Formula 2] is performed to create a profile that is more continuous.

Therefore, if the corrected raw data R ′ satisfying the following [Expression 3] and the first expression such that the Sum value of the v-view angle coincides with the above smoothing result can be calculated, the data is corrected based on this. I can do it. Since noise is mixed not in raw data but in the X-ray dose, in the first embodiment of the present invention, correction is performed on the X-ray dose before conversion. That is, a noise correction constant n for each view angle that satisfies the second equation of [Equation 3] is found, and the raw data R ′ is obtained.

  However, the calculation of the noise correction constant n at each view angle satisfying the second expression of [Equation 3] is not easy because the calculation for obtaining the corrected raw data R 'may be a logarithmic function. Furthermore, since the X-ray dose data before conversion is corrected and corrected raw data R ′ is calculated, the conversion to raw data is eventually performed twice. In the first embodiment, corrected raw data is It takes a long processing time to find out.

Therefore, in the second embodiment of the present invention, the steps of inverse conversion from raw data to X-ray dose in advance, correction by a noise correction constant n ₀ common to the target apparatus, and re-conversion from corrected X-ray dose to raw data. As shown in FIG. 5, the processing time is increased by formulating [Equation 4] for the correction amount ΔR of raw data that assumes and reflects the above.

By using the table of FIG. 5, the correction amount ΔR of R → R ′ with respect to the noise correction constant n _{0 is} extracted from the table, and the corrected raw data R ′ can be easily obtained by using [Equation 6] described below. it can. As for the noise correction constant n ₀ , an appropriate one is selected by experimentally measuring the target device in advance. For example, 1/2 of the noise amplitude is selected. If the memory capacity for tabulation is not sufficient and the accuracy cannot be ensured, the accuracy can be maintained by calculating the correction amount ΔR by interpolation as shown in FIG. First, the correction amount ΔR of each channel (detection element) is obtained for each view angle using this table, and the total value CorrSum is calculated as shown in [Equation 5].

Next, using the calculated CorrSum, corrected raw data R ′ is obtained according to [Equation 6].

Sum value of the above corrected raw data R '
As can be seen from [Equation 7] below, it can be seen that the profile is equal to Smooth and has a continuous Sum value.

Therefore, in the second embodiment, by defining a noise correction constant n ₀ common to the target apparatus, it is possible to obtain [Equation 3] without finding the noise correction constant n for each view angle as in the first embodiment. The correction raw data R ′ satisfying the condition of the first expression can be easily obtained using a table prepared in advance.

A flowchart for calculating the corrected raw data R ′ according to the second embodiment described above is shown in FIG.
That is, first, in step 61, the Sum value of the raw data R for each view angle
Is calculated. Next, in step 62, it is determined whether the above processing has been executed for all the view angles. If No, step 61 is repeated, and if Yes, the process proceeds to step 63. Here, smoothing for each view angle is performed according to [Equation 2]. Data Smooth is calculated. Next, in step 64, a raw data correction amount ΔR for each channel (detection element) is first calculated using a correction table prepared in advance for the target device as shown in FIG. A total value CorrSum of the correction amount ΔR shown is calculated. Next, in step 65, corrected raw data R ′ for each channel (detection element) is calculated using the Sum value, Smooth value, correction amount ΔR and CorrSum value obtained in steps 61, 63 and 64, respectively. Finally, in step 66, it is determined whether or not the processing of steps 64 and 65 has been executed for all view angles. If No, steps 64 and 65 are repeated, and if Yes, the raw data correction processing is terminated.

  FIG. 7A shows a tomographic image of a shoulder portion by a conventional apparatus that is not corrected to raw data, and FIG. 7B shows the same portion reconstructed using raw data corrected by the present invention. An image is shown. This shows that an image with less noise can be created according to the present invention.

An example of the X-ray CT image of the shoulder part containing an artifact. (A) is a figure explaining the outline of the data collection by X-ray CT apparatus. (B) is a figure which shows an example of the sinogram produced | generated from the data collected in the image processing apparatus of X-ray CT apparatus. The flow figure explaining the creation procedure of raw data. The figure explaining the smoothing about the Sum value for every view angle performed by this invention. An example of the table for calculating | requiring the variation | change_quantity (DELTA) R of raw data utilized in 2nd Example of this invention. The flowchart explaining the correction | amendment raw data creation procedure by 2nd Example of this invention. (A) is an example of an X-ray CT tomographic image of a shoulder by a conventional apparatus that is not corrected to raw data similar to FIG. (B) is an example of an X-ray tomographic image of the same portion reconstructed using the corrected raw data according to the present invention.

Explanation of symbols

21 ... X-ray source
22 ... Subject
23 ... Detector
50. Image processing device

Claims (3)

A detection comprising an X-ray source that can be rotated around the subject and arranged opposite to the subject, and a number of detection elements that detect the X-rays emitted from the X-ray source and transmitted through the subject. In the X-ray CT apparatus having an imager and an image processing apparatus for imaging the detected X-ray dose , the image processing apparatus performs logarithmic conversion on projection X-ray dose detected by each detection element for each view angle. Means for calculating the sum value Sum of raw data;
Means for smoothing the calculated raw data sum Sum for each view angle along the view angle to obtain a smoothed raw data sum Smooth;
An X-ray CT apparatus comprising means for obtaining corrected raw data R ′ for each detection element at each view angle so that the raw data total value Sum at each view angle becomes a smoothed raw data total value Smooth .   The image processing apparatus obtains a noise correction constant n for the transmitted X-ray dose detected for each view angle, and a noise correction constant n for obtaining the correction raw data for each detection element for each view angle. Means for correcting the detected transmitted X-ray dose, and means for obtaining corrected raw data for each detection element at each view angle based on the corrected transmitted X-ray dose for each detection element at each view angle. The X-ray CT apparatus according to claim 1, comprising: The image processing apparatus is created using a noise correction constant n ₀ that is determined in common between the elements for the detected transmitted X-rays when obtaining correction raw data for each detection element at each view angle. 2. The X-ray CT apparatus according to claim 1, further comprising means for obtaining corrected raw data for each detection element at each view angle based on a map of the raw data and the corresponding raw data correction amount.