Image processing device, image processing system, image processing method and program

Abstract

To acquire an image with reduced noise.SOLUTION: An image processing device includes a processing part for executing noise reduction processing and material separation processing by using a plurality of images corresponding to a plurality of mutually-different radiation energies acquired by irradiating a subject with radiation and capturing an image. The processing part generates a noise-reduced image acquired by applying a material thickness image acquired by the material separation processing to filter processing for every frequency component decomposed into multiple frequencies, and generates an image by the material separation processing performed again by using the noise-reduced thickness image formed by combining the noise-reduced images for every frequency component, and an accumulated image acquired on the basis of addition of the plurality of images.SELECTED DRAWING: Figure 16

Description

[Technical Field]

[0001] The disclosed technology relates to an image processing device, a radiation imaging system, an image processing method, and a program. More specifically, the present invention relates to an image processing device, a radiation imaging system, an image processing method, and a program used for still image capture such as general radiography in medical diagnosis and for moving image capture such as fluoroscopic radiography. [Background technology]

[0002] Currently, radiation imaging devices using flat panel detectors (hereinafter abbreviated as FPD) made of semiconductor materials are widely used as imaging devices for medical image diagnosis and non-destructive testing using X-rays.

[0003] Energy subtraction, an imaging method using an FPD, processes multiple images of different energies obtained by irradiating X-rays with different tube voltages, making it possible to obtain material-separated images with reduced contrast, such as bone images or soft tissue images (Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2019-162358 Summary of the Invention [Problem to be solved by the invention]

[0005] In interventional radiology (IVR) using FPD, contrast agents are injected into blood vessels, and medical devices such as catheters and guidewires are inserted into the blood vessels, and treatment is performed while checking the position and shape of the contrast agents and medical devices. When energy subtraction is used to reduce the contrast of soft tissue and bone, the visibility of the contrast agents and medical devices improves, but the problem of increased image noise can arise.

[0006] The disclosed technology provides a technology that makes it possible to acquire an image with reduced noise. [Means for solving the problem]

[0007] An image processing device according to one aspect of the disclosed technique includes a processing means for performing noise reduction processing and material separation processing using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and performing imaging; The processing means generating a noise-reduced image by applying a filter process to each frequency component obtained by decomposing the thickness image of the material obtained by the material separation process into a plurality of frequencies; An image is generated by performing a second material separation process using a noise-reduced thickness image obtained by combining the noise-reduced images for each frequency component and an accumulated image obtained by adding the multiple images. [Effects of the Invention]

[0008] According to the disclosed technology, it is possible to obtain an image with reduced noise. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram showing an example of the configuration of an X-ray imaging system according to a first embodiment. [Figure 2] FIG. 2 is a pixel equivalent circuit diagram of the X-ray imaging device according to the first embodiment. [Figure 3] 4 is a timing chart of the X-ray imaging apparatus according to the first embodiment. [Figure 4] 4 is a timing chart of the X-ray imaging apparatus according to the first embodiment. [Figure 5] 5A to 5C are diagrams illustrating correction processing according to the first embodiment. [Figure 6] FIG. 2 is a block diagram of signal processing according to the first embodiment. [Figure 7] FIG. 2 is a block diagram of image processing according to the first embodiment. [Figure 8]FIG. 3 is a block diagram of a correction process according to the first embodiment. [Figure 9] 3A and 3B are diagrams illustrating an example of a file image A and a bone image B according to the first embodiment. [Figure 10] FIG. 2 is a block diagram of signal processing in the radiation imaging system according to the first embodiment. [Figure 11] 10A and 10B are diagrams illustrating noise reduction processing using a Gaussian filter. [Figure 12] 10A and 10B are diagrams illustrating noise reduction processing using an epsilon filter. [Figure 13] 10A and 10B are diagrams for explaining an example of edge determination processing using a stored image. [Figure 14] 10A to 10C are diagrams for explaining an example of noise reduction processing by frequency decomposition according to the second embodiment. [Figure 15] 10A and 10B are diagrams for explaining an example of noise reduction processing using frequency decomposition and a recursive filter according to the second embodiment. [Figure 16] 13A and 13B are diagrams illustrating an example of a configuration in which edge determination using a stored image is applied in noise reduction processing using an epsilon filter according to the third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention claimed. Although multiple features are described in the embodiments, not all of these multiple features are necessarily essential to the invention, and multiple features may be combined arbitrarily. Furthermore, in the accompanying drawings, the same reference numerals are used to designate the same or similar components, and redundant explanations will be omitted.

[0011] Note that radiation in the disclosed technology includes α-rays, β-rays, γ-rays, and the like, which are beams created by particles (including photons) emitted by radioactive decay, as well as beams having the same or higher energy levels, such as X-rays, particle beams, and cosmic rays. In the following embodiments, an apparatus using X-rays as an example of radiation will be described. Therefore, hereinafter, a radiation imaging apparatus and a radiation imaging system will be described as an X-ray imaging apparatus and an X-ray imaging system, respectively.

[0012] (First embodiment) 1 is a block diagram showing an example of the configuration of an X-ray imaging system as an example of a radiation imaging system according to Embodiment 1. The X-ray imaging system of the first embodiment includes an X-ray generation device 101, an X-ray control device 102, an imaging control device 103, and an X-ray imaging device 104 (radiation imaging device).

[0013] The X-ray generator 101 generates X-rays and irradiates the subject with the X-rays. The X-ray controller 102 controls the generation of X-rays in the X-ray generator 101. The imaging controller 103 has, for example, one or more processors (CPUs) and memory, and the processor executes a program stored in the memory to acquire X-ray images and perform image processing. Note that each process, including image processing by the imaging controller 103, may be realized by dedicated hardware or by a combination of hardware and software. The X-ray imaging device 104 has a phosphor 105 that converts X-rays into visible light and a two-dimensional detector 106 that detects the visible light. The two-dimensional detector is a sensor in which pixels 20 that detect X-ray quanta are arranged in an array of X columns by Y rows, and outputs image information.

[0014] The imaging control device 103 functions as an image processing device that processes radiation images using the above-mentioned processor. The acquisition unit 131, correction unit 132, signal processing unit 133, and image processing unit 134 show an example of the functional configuration of the image processing device. The radiation imaging system includes an X-ray imaging device 104 (radiation imaging device) communicably connected to the imaging control device 103 (image processing device).

[0015] The acquisition unit 131 acquires a plurality of radiological images having different energies obtained by irradiating a subject with radiation and capturing the image. The acquisition unit 131 acquires the plurality of radiological images by performing sample and hold multiple times during the irradiation of one shot of radiation.

[0016] The correction unit 132 corrects the multiple radiation images acquired by the acquisition unit 131 to generate multiple images to be used in the energy subtraction process.

[0017] The signal processing unit 133 performs noise reduction processing and material separation processing by energy subtraction using multiple images corresponding to multiple different radiation energies obtained by irradiating the subject with radiation and performing imaging. The signal processing unit 133 generates noise-reduced images by applying filter processing to each frequency component obtained by decomposing the material thickness image obtained by the material separation processing into multiple frequencies, and generates an image by performing material separation processing again using a noise-reduced thickness image obtained by combining the noise-reduced images for each frequency component and an accumulated image obtained by adding the multiple images. The signal processing unit 133 generates material thickness images using multiple images corresponding to multiple different radiation energies obtained by irradiating the subject with radiation and performing imaging. The signal processing unit 133 generates noise-reduced images for each frequency component by applying filter processing to the material thickness image, and generates a noise-reduced thickness image by combining the noise-reduced images. The signal processing unit 133 generates a material characteristic image using the multiple images generated by the correction unit 132. A material characteristic image is an image acquired in an energy subtraction process, such as a material separation image that shows materials separately, such as bone and soft tissue, or a material identification image that shows effective atomic number and its surface density.

[0018] The multiple radiation energies include a first radiation energy (high energy) and a second radiation energy (low energy) lower than the first radiation energy. The signal processing unit 133 performs material separation processing using a first image (high energy image H) captured with the first radiation energy (high energy) and a second image (low energy image L) captured with the second radiation energy (low energy) to generate a first material separated image showing the thickness of a first material contained in the subject and a second material separated image showing the thickness of a second material different from the first material. The signal processing unit 133 performs material separation processing using the multiple radiation images captured with different radiation energies to generate a first material separated image showing the thickness of the first material and a second material separated image showing the thickness of a second material different from the first material. The signal processing unit 133 also generates a thickness image combining the thickness of the first material and the thickness of the second material. Here, the first material includes at least calcium, hydroxyapatite, or bone, and the second material includes at least water, fat, or soft material not containing calcium. The signal processing unit 133 will be described in detail later. The image processing unit 134 uses the material property image acquired by the signal processing performed by the signal processing unit 133 to generate an image for display.

[0019] 2 is an equivalent circuit diagram of a pixel 20 according to the first embodiment. The pixel 20 includes a photoelectric conversion element 201 and an output circuit unit 202. The photoelectric conversion element 201 can typically be a photodiode. The output circuit unit 202 includes an amplifier circuit unit 204, a clamp circuit unit 206, a sample-and-hold circuit 207, and a selection circuit unit 208.

[0020] The photoelectric conversion element 201 includes a charge accumulation section, which is connected to the gate of a MOS transistor 204a in the amplifier circuit section 204. The source of the MOS transistor 204a is connected to a current source 204c via a MOS transistor 204b. The MOS transistor 204a and the current source 204c form a source follower circuit. The MOS transistor 204b is an enable switch that turns on when an enable signal EN supplied to its gate reaches an active level, putting the source follower circuit into an operating state.

[0021] 2, the charge storage section of the photoelectric conversion element 201 and the gate of the MOS transistor 204a form a common node, and this node functions as a charge-voltage converter that converts the charge stored in the charge storage section into a voltage. That is, a voltage V (=Q / C) appears in the charge-voltage converter, which is determined by the charge Q stored in the charge storage section and the capacitance C of the charge-voltage converter. The charge-voltage converter is connected to a reset potential Vres via a reset switch 203. When the reset signal PRES becomes active level, the reset switch 203 turns on, and the potential of the charge-voltage converter is reset to the reset potential Vres.

[0022] The clamp circuit unit 206 clamps noise output by the amplifier circuit unit 204 in response to the reset potential of the charge-voltage converter using a clamp capacitor 206a. In other words, the clamp circuit unit 206 is a circuit for canceling this noise from the signal output from the source follower circuit in response to the charge generated by photoelectric conversion in the photoelectric conversion element 201. This noise includes kTC noise during reset. Clamping is achieved by first setting the clamp signal PCL to an active level to turn on the MOS transistor 206b, and then setting the clamp signal PCL to an inactive level to turn off the MOS transistor 206b. The output side of the clamp capacitor 206a is connected to the gate of the MOS transistor 206c. The source of the MOS transistor 206c is connected to a current source 206e via a MOS transistor 206d. The MOS transistor 206c and the current source 206e form a source follower circuit. The MOS transistor 206d is an enable switch that turns on when an enable signal EN0 supplied to its gate becomes active, activating the source follower circuit.

[0023] The signal output from the clamp circuit unit 206 in response to the charge generated by photoelectric conversion in the photoelectric conversion element 201 is written as an optical signal into the capacitor 207Sb via the switch 207Sa when the optical signal sampling signal TS becomes active. The signal output from the clamp circuit unit 206 when the MOS transistor 206b is turned on immediately after resetting the potential of the charge-voltage converter is a clamp voltage. The noise signal is written into the capacitor 207Nb via the switch 207Na when the noise sampling signal TN becomes active. This noise signal includes an offset component of the clamp circuit unit 206. The switch 207Sa and the capacitor 207Sb form a signal sample-and-hold circuit 207S, and the switch 207Na and the capacitor 207Nb form a noise sample-and-hold circuit 207N. The sample-and-hold circuit unit 207 includes the signal sample-and-hold circuit 207S and the noise sample-and-hold circuit 207N.

[0024] When the driving circuit unit drives the row selection signal to an active level, the signal (optical signal) held in the capacitor 207Sb is output to the signal line 21S via the MOS transistor 208Sa and the row selection switch 208Sb. At the same time, the signal (noise) held in the capacitor 207Nb is output to the signal line 21N via the MOS transistor 208Na and the row selection switch 208Nb. The MOS transistor 208Sa forms a source follower circuit together with a constant current source (not shown) provided on the signal line 21S. Similarly, the MOS transistor 208Na forms a source follower circuit together with a constant current source (not shown) provided on the signal line 21N. The MOS transistor 208Sa and the row selection switch 208Sb form the signal selection circuit unit 208S, and the MOS transistor 208Na and the row selection switch 208Nb form the noise selection circuit unit 208N. The selection circuit section 208 includes a signal selection circuit section 208S and a noise selection circuit section 208N.

[0025] The pixel 20 may have an addition switch 209S that adds optical signals from multiple adjacent pixels 20. In addition mode, an addition mode signal ADD becomes active, and the addition switch 209S is turned on. As a result, the capacitances 207Sb of adjacent pixels 20 are connected to each other by the addition switch 209S, and the optical signals are averaged. Similarly, the pixel 20 may have an addition switch 209N that adds noises from multiple adjacent pixels 20. When the addition switch 209N is turned on, the capacitances 207Nb of adjacent pixels 20 are connected to each other by the addition switch 209N, and the noises are averaged. The adder 209 includes an addition switch 209S and an addition switch 209N.

[0026] The pixel 20 may also have a sensitivity change unit 205 for changing the sensitivity. The pixel 20 may include, for example, a first sensitivity change switch 205a and a second sensitivity change switch 205'a, as well as associated circuit elements. When the first change signal WIDE becomes active, the first sensitivity change switch 205a is turned on, and the capacitance of the first additional capacitor 205b is added to the capacitance of the charge-voltage conversion unit. This reduces the sensitivity of the pixel 20. When the second change signal WIDE2 becomes active, the second sensitivity change switch 205'a is turned on, and the capacitance of the second additional capacitor 205'b is added to the capacitance of the charge-voltage conversion unit. This further reduces the sensitivity of the pixel 20. By adding a function to reduce the sensitivity of the pixel 20 in this way, it becomes possible to receive a larger amount of light and widen the dynamic range. When the first change signal WIDE becomes active, the enable signal ENw may be set to active, causing the MOS transistor 204'a to operate as a source follower instead of the MOS transistor 204a.

[0027] The X-ray imaging device 104 reads the output of the pixel circuit as described above from the two-dimensional detector 106 , converts it into a digital value using an AD converter (not shown), and then transfers the image to the imaging control device 103 .

[0028] Next, the operation of the X-ray imaging system of the first embodiment having the above-mentioned configuration will be described. Fig. 3 shows the drive timing of the X-ray imaging device 104 when energy subtraction is performed in the X-ray imaging system of the first embodiment. The waveform in Fig. 3, with time on the horizontal axis, shows the timing of X-ray exposure, synchronization signal, resetting of the photoelectric conversion element 201, sample and hold circuit 207, and readout of an image from the signal line 21.

[0029] X-rays are emitted after the photoelectric conversion element 201 is reset by a reset signal. Ideally, the X-ray tube voltage is a square wave, but it takes a finite amount of time for the tube voltage to rise and fall. In particular, when pulsed X-rays are used and the exposure time is short, the tube voltage can no longer be considered a square wave, and takes on a waveform such as that shown by X-rays 301 to 303. The X-rays 301 in the rising phase, X-rays 302 in the stable phase, and X-rays 303 in the falling phase all have different X-ray energies. Therefore, by obtaining X-ray images corresponding to the radiation in the periods separated by sample and hold, multiple types of X-ray images with different energies can be obtained.

[0030] In the X-ray imaging device 104, after the X-rays 301 in the rising phase are irradiated, sampling is performed by the noise sample and hold circuit 207N, and after the X-rays 302 in the stable phase are irradiated, sampling is performed by the signal sample and hold circuit 207S. After that, the X-ray imaging device 104 reads out the difference between the signal lines 21N and 21S as an image. At this time, the noise sample and hold circuit 207N holds the signal (R1) of the X-rays 301 in the rising phase, and the signal sample and hold circuit 207S holds the sum (R1+B) of the signal of the X-rays 301 in the rising phase and the signal (B) of the X-rays 302 in the stable phase. Therefore, an image 304 corresponding to the signal of the X-rays 302 in the stable phase is read out.

[0031] Next, after the X-ray imaging device 104 completes the irradiation of X-rays 303 in the falling edge period and the readout of image 304, it again samples using the signal sample-and-hold circuit 207S. Thereafter, the X-ray imaging device 104 resets the photoelectric conversion element 201, samples again using the noise sample-and-hold circuit 207N, and reads out the difference between signal line 21N and signal line 21S as an image. At this time, the noise sample-and-hold circuit 207N holds a signal in a state where no X-rays are irradiated, and the signal sample-and-hold circuit 207S holds the sum (R1+B+R2) of the signal of X-rays 301 in the rising edge period, the signal of X-rays 302 in the stable edge period, and the signal (R2) of X-rays 303 in the falling edge period. Therefore, an image 306 corresponding to the signal of X-rays 301 in the rising edge period, the signal of X-rays 302 in the stable edge period, and the signal of X-rays 303 in the falling edge period is read out. Thereafter, by calculating the difference between image 306 and image 304, an image 305 corresponding to the sum of X-rays 301 in the rising phase and X-rays 303 in the falling phase is obtained. This calculation may be performed by the X-ray imaging device 104 or the imaging control device 103.

[0032] The timing for resetting the sample-and-hold circuit 207 and the photoelectric conversion element 201 is determined using a synchronization signal 307 that indicates the start of X-ray exposure from the X-ray generator 101. A method for detecting the start of X-ray exposure may be, but is not limited to, a configuration in which the tube current of the X-ray generator 101 is measured and whether or not the current value exceeds a preset threshold. For example, a configuration in which, after the resetting of the photoelectric conversion element 201 is completed, pixels 20 are repeatedly read out and whether or not the pixel value exceeds a preset threshold may be used to detect the start of X-ray exposure.

[0033] Alternatively, for example, a configuration may be used in which an X-ray detector different from the two-dimensional detector 106 is built into the X-ray imaging device 104, and the start of X-ray exposure is detected by determining whether or not the measurement value of the X-ray detector exceeds a preset threshold. In either method, sampling by the signal sample-and-hold circuit 207S, sampling by the noise sample-and-hold circuit 207N, and resetting of the photoelectric conversion element 201 are performed after a predetermined time has elapsed since the input of a synchronization signal 307 indicating the start of X-ray exposure.

[0034] In this way, an image 304 corresponding to the stable period of the pulsed X-rays and an image 305 corresponding to the sum of the rising and falling periods are obtained. Because the energies of the X-rays irradiated when forming these two X-ray images are different, energy subtraction processing can be performed by performing calculations between these X-ray images.

[0035] Fig. 4 shows the drive timing of the X-ray imaging device 104 when energy subtraction is performed in the X-ray imaging system according to the first embodiment. The drive timing shown in Fig. 4 differs from the drive timing shown in Fig. 3 in that the tube voltage of the X-ray generator 101 is actively switched.

[0036] First, after the photoelectric conversion element 201 is reset, the X-ray generation device 101 radiates low-energy X-rays 401. In this state, the X-ray imaging device 104 performs sampling using the noise sample-and-hold circuit 207N. Thereafter, the X-ray generation device 101 switches the tube voltage to radiate high-energy X-rays 402. In this state, the X-ray imaging device 104 performs sampling using the signal sample-and-hold circuit 207S. Thereafter, the X-ray generation device 101 switches the tube voltage to radiate low-energy X-rays 403. The X-ray imaging device 104 reads out the difference between the signal lines 21N and 21S as an image. At this time, the noise sample-and-hold circuit 207N holds the signal (R1) of the low-energy X-rays 401, and the signal sample-and-hold circuit 207S holds the sum (R1+B) of the signal of the low-energy X-rays 401 and the signal (B) of the high-energy X-rays 402. Therefore, an image 404 corresponding to the signal of the high-energy X-rays 402 is read out.

[0037] Next, after the X-ray imaging device 104 completes the irradiation of low-energy X-rays 403 and the readout of image 404, it again samples using the signal sample-and-hold circuit 207S. Thereafter, the X-ray imaging device 104 resets the photoelectric conversion element 201, and again samples using the noise sample-and-hold circuit 207N, reading out the difference between signal line 21N and signal line 21S as an image. At this time, the noise sample-and-hold circuit 207N holds a signal in a state where no X-rays are irradiated, and the signal sample-and-hold circuit 207S holds the sum (R1+B+R2) of the signal of low-energy X-rays 401, the signal of high-energy X-rays 402, and the signal (R2) of low-energy X-rays 403. Therefore, an image 406 corresponding to the signal of low-energy X-rays 401, the signal of high-energy X-rays 402, and the signal of low-energy X-rays 403 is read out.

[0038] Thereafter, by calculating the difference between image 406 and image 404, an image 405 corresponding to the sum of low-energy X-rays 401 and low-energy X-rays 403 is obtained. This calculation may be performed by the X-ray imaging device 104 or the imaging control device 103. The synchronization signal 407 is the same as in FIG. 3. In this way, by acquiring images while actively switching the tube voltage, it is possible to make the energy difference between low-energy and high-energy radiation images larger than in the method of FIG. 3.

[0039] Next, a description will be given of the energy subtraction processing performed by the imaging control device 103. The energy subtraction processing in the first embodiment is divided into three stages: correction processing by the correction unit 132, signal processing by the signal processing unit 133, and image processing by the image processing unit 134. Each of these processes will be described below.

[0040] The correction process is a process of processing a plurality of radiation images acquired from the X-ray imaging device 104 to generate a plurality of images to be used in the signal processing described later in the energy subtraction process. FIG. 5 shows a block diagram of the correction process for the energy subtraction process according to the first embodiment. First, the acquisition unit 131 causes the X-ray imaging device 104 to perform imaging without emitting X-rays, and acquires images using the drive shown in FIG. 3 or FIG. 4. Two images are read out by this drive. Hereinafter, the first image (image 304 or image 404) is referred to as F_ODD, and the second image (image 306 or image 406) is referred to as F_EVEN. F_ODD and F_EVEN are images corresponding to fixed pattern noise (FPN) of the X-ray imaging device 104.

[0041] Next, the acquisition unit 131 causes the X-ray imaging device 104 to irradiate X-rays in an absence of an object and perform imaging, and acquires images for gain correction output from the X-ray imaging device 104 by the driving shown in FIG. 3 or FIG. 4. This driving reads out two images in the same manner as above. Hereinafter, the first image for gain correction (image 304 or image 404) is referred to as W_ODD, and the second image for gain correction (image 306 or image 406) is referred to as W_EVEN. W_ODD and W_EVEN are images corresponding to the sum of the FPN of the X-ray imaging device 104 and the signal due to X-rays. The correction unit 132 subtracts F_ODD from W_ODD and F_EVEN from W_EVEN to obtain images WF_ODD and WF_EVEN from which the FPN of the X-ray imaging device 104 has been removed. This is called offset correction.

[0042] WF_ODD is an image corresponding to X-rays 302 in the stable period, and WF_EVEN is an image corresponding to the sum of X-rays 301 in the rising period, X-rays 302 in the stable period, and X-rays 303 in the falling period. Therefore, by subtracting WF_ODD from WF_EVEN, the correction unit 132 obtains an image corresponding to the sum of X-rays 301 in the rising period and X-rays 303 in the falling period. This process of subtracting multiple images to obtain an image corresponding to X-rays in a specific period separated by sample and hold is called color correction. The energy of X-rays 301 in the rising period and X-rays 303 in the falling period is lower than the energy of X-rays 302 in the stable period. Therefore, by subtracting WF_ODD from WF_EVEN through color correction, a low-energy image W_Low when there is no subject is obtained. Furthermore, a high-energy image W_High when there is no subject is obtained from WF_ODD.

[0043] Next, the acquisition unit 131 causes the X-ray imaging device 104 to irradiate X-rays and perform imaging when an object is present, and acquires images output from the X-ray imaging device 104 by the driving shown in FIG. 3 or 4. At this time, two images are read out. Hereinafter, the first image (image 304 or image 404) is referred to as X_ODD, and the second image (image 306 or image 406) is referred to as X_EVEN. The correction unit 132 performs offset correction and color correction similar to those performed when an object is not present, thereby obtaining a low-energy image X_Low when an object is present and a high-energy image X_High when an object is present.

[0044] Here, if the thickness of the subject is d, the linear attenuation coefficient of the subject is μ, the output of pixel 20 when there is no subject is I0, and the output of pixel 20 when there is a subject is I, the following equation (1) holds.

[0045]

number

[0046] Transforming equation (1) yields equation (2) below. The right-hand side of equation (2) represents the attenuation rate of the subject. The attenuation rate of the subject is a real number between 0 and 1.

[0047]

number

[0048] Therefore, the correction unit 132 obtains an image L of the attenuation ratio at low energy (hereinafter also referred to as "low-energy image L") by dividing the low-energy image X_Low when a subject is present by the low-energy image W_Low when no subject is present. Similarly, the correction unit 132 obtains an image H of the attenuation ratio at high energy (hereinafter also referred to as "high-energy image H") by dividing the high-energy image X_High when a subject is present by the high-energy image W_High when no subject is present. The process of obtaining an image (L, H) of the attenuation ratio at low energy or the attenuation ratio at high energy by dividing an image obtained based on a radiographic image obtained in the presence of a subject by an image obtained based on a radiographic image obtained in the absence of a subject is called gain correction.

[0049] Fig. 6 shows a block diagram of signal processing for energy subtraction processing according to the first embodiment. The signal processing unit 133 generates a material characteristic image using multiple images obtained from the correction unit 132. Below, a description will be given of processing for generating a material separation image consisting of a bone thickness image B (hereinafter also referred to as bone image B) and a soft tissue thickness image S (hereinafter also referred to as soft tissue image S). The signal processing unit 133 obtains the bone thickness image B and the soft tissue thickness image S from the image L of the attenuation ratio at low energy and the image H of the attenuation ratio at high energy obtained by the correction shown in Fig. 5 through the following processing.

[0050] First, let E be the energy of the X-ray photon, N(E) be the number of photons at energy E, B be the thickness in the bone thickness image, S be the thickness in the soft tissue thickness image, and μ be the linear attenuation coefficient of bone at energy E. B (E), the linear attenuation coefficient of soft tissue at energy E is μ S (E), if the attenuation rate is I / I0, the following equation [Equation 3] holds.

[0051]

number

[0052] The number of photons N(E) at energy E is the X-ray spectrum. The X-ray spectrum can be obtained by simulation or actual measurement. The linear attenuation coefficient μ of bone at energy E is B (E) and the linear attenuation coefficient μ of soft tissue at energy E S (E) can be obtained from databases such as those of the National Institute of Standards and Technology (NIST). Therefore, according to Equation 3, it is possible to calculate the thickness B in an arbitrary bone thickness image, the thickness S in an arbitrary soft tissue thickness image, and the attenuation ratio I / I0 in the X-ray spectrum N(E).

[0053] Here, the spectrum of low energy X-rays is N L (E) High-energy X-ray spectrum H Assuming (E), the following equations in [Equation 4] hold for the attenuation rate in image L and the attenuation rate in image H. In the following explanation, the attenuation rate in image L shown in [Equation 4] will also be referred to simply as the low-energy attenuation rate L, and the attenuation rate in image H will also be referred to simply as the high-energy attenuation rate H.

[0054]

number

[0055] By solving the nonlinear simultaneous equations in [Equation 4], the thickness B in the bone thickness image and the thickness S in the soft tissue thickness image can be obtained. As a typical method for solving the nonlinear simultaneous equations, the Newton-Raphson method will be described here. First, let us set the number of iterations of the Newton-Raphson method to m, and let B be the thickness of the bone after the mth iteration. m , the thickness of the soft tissue after the mth iteration is S m Then, the attenuation rate of high energy after the mth iteration, H m , the low-energy attenuation rate L after the mth iteration m is expressed by the following equation [5].

[0056]

number

[0057] The rate of change in attenuation rate when the thickness changes slightly is expressed by the following formula [6].

[0058]

number

[0059] At this time, the bone thickness B after the m+1th iteration m+1 and soft tissue thickness S m+1 is expressed by the following equation (7) using the attenuation rate of high energy H and the attenuation rate of low energy L.

[0060]

number

[0061] The inverse matrix of a 2x2 matrix, with its determinant det, is expressed by the following equation (8) using Cramer's rule.

[0062]

number

[0063] Therefore, by substituting the formula [8] into the formula [7], the following formula [9] is obtained.

[0064]

number

[0065] By repeating the above calculations, the attenuation rate of high energy after the mth iteration, H m The difference between the measured high-energy attenuation rate H and the measured high-energy attenuation rate L approaches 0. The same is true for the low-energy attenuation rate L. As a result, the bone thickness B after the mth iteration m converges to the bone thickness B, and the m-th soft tissue thickness S mconverges to the thickness S of the soft tissue. In this way, the nonlinear simultaneous equations shown in Equation 4 can be solved. Therefore, by calculating Equation 4 for all pixels, it is possible to obtain an image B of the bone thickness and an image S of the soft tissue thickness from the image L of the attenuation rate at low energy and the image H of the attenuation rate at high energy.

[0066] In the first embodiment, an image B of bone thickness and an image S of soft tissue thickness are calculated, but the disclosed technology is not limited to this. For example, the thickness W of water and the thickness I of contrast agent may be calculated. In other words, they may be decomposed into the thicknesses of any two types of material. Furthermore, an image of effective atomic number Z and an image of surface density D may be obtained from an image L of attenuation rate at low energy and an image H of attenuation rate at high energy obtained by the correction shown in FIG. 5. The effective atomic number Z is the equivalent atomic number of the mixture, and the surface density D is the density [g / cm] of the subject. 3 ] and the thickness of the subject [cm].

[0067] In the first embodiment, the nonlinear simultaneous equations are solved using the Newton-Raphson method. However, the disclosed technology is not limited to this configuration. For example, an iterative solution method such as the least squares method or bisection method may be used. In the first embodiment, the nonlinear simultaneous equations are solved using an iterative solution method, but the disclosed technology is not limited to this configuration. A configuration may be used in which bone thicknesses B and soft tissue thicknesses S for various combinations of high-energy attenuation rates H and low-energy attenuation rates L are calculated in advance to generate a table, and the bone thicknesses B and soft tissue thicknesses S are quickly calculated by referring to this table.

[0068] FIG. 7 shows a block diagram of image processing in energy subtraction processing according to the first embodiment. The image processing unit 134 according to the first embodiment generates an image for display by performing post-processing on the bone thickness image B obtained by the signal processing shown in FIG. 6. As post-processing, the image processing unit 134 can use logarithmic transformation, dynamic range compression, etc. By inputting the type and strength of post-processing as parameters, the image processing unit 134 can switch the content of processing.

[0069] FIG. 8 shows a block diagram of the correction process according to the first embodiment. The display image in FIG. 7 can be an image without energy resolution, i.e., an accumulated image A that is compatible with images captured by an existing radiation imaging system. The signal processing unit 133 can generate accumulated image A through the following process. For example, as shown in FIG. 8, accumulated image A can be generated by dividing image XF_EVEN, which corresponds to the sum of the rising period 301, stable period 302, and falling period 303 of X-rays when a subject is present, by image WF_EVEN, which corresponds to the sum of the rising period 301, stable period 302, and falling period 303 of X-rays when no subject is present. Alternatively, accumulated image A can be obtained by adding together multiple images corresponding to different radiation energies. For example, accumulated image A can be generated by multiplying image H, which shows the attenuation rate at high energy, and image L, which shows the attenuation rate at low energy, by a coefficient and then adding them together. For example, in calculating accumulated image A, one coefficient can be set to 0 and the other coefficient to 1, and image H or image L itself can be used as accumulated image A. Alternatively, a stored image may be generated by multiplying at least one of a plurality of images by a coefficient and adding the resulting images together. An image that has been captured at substantially the same time as the image to be subjected to energy subtraction processing but has not undergone energy subtraction processing may be used as stored image A.

[0070] [Number 10] A=XF_EVEN / WF_EVEN =X_High / WF_EVEN+X_Low / WF_EVEN =H*(W_High / WF_EVEN)+L*(W_Low / WF_EVEN) FIG. 9 is a diagram illustrating an example of a stored image A and a bone image B according to the first embodiment. A normal human body is composed only of soft tissue and bone. However, when performing interventional radiology (IVR) using the radiation imaging system shown in FIG. 1, a contrast agent is injected into blood vessels. Furthermore, a catheter or guide wire is inserted into the blood vessel, and a procedure such as placing a stent or coil is performed. IVR is performed while checking the position and shape of the contrast agent or medical device. Therefore, visibility may be improved by isolating only the contrast agent or medical device, or by removing background components such as soft tissue or bone.

[0071] 9, in an image compatible with a normal radiation imaging system, i.e., stored image A, soft tissue is displayed in addition to the contrast agent, stent, and bone. On the other hand, in bone image B obtained by the radiation imaging system according to the first embodiment, the contrast of the soft tissue can be reduced.

[0072] On the other hand, the main component of the contrast agent is iodine, and the main component of the medical device is a metal such as stainless steel. Both have a higher atomic number than calcium, the main component of bone, so bone, contrast agent, and medical device are displayed in bone image B.

[0073] The inventors of the present application conducted research and found that even when separating a water image W and a contrast agent image I based on a high-energy image H and a low-energy image L, the bone, contrast agent, and medical device were still displayed in the contrast agent image I. The same was true when the tube voltages and filters for the low-energy X-rays and the high-energy X-rays were changed. In either case, the bone, contrast agent, and medical device were displayed in the bone image B. In other words, it was not possible to separate the contrast agent and medical device from the bone image B.

[0074] That is, when the contrast of soft tissue reduces visibility, such as in the lung field during chest IVR, displaying bone image B in the radiation imaging system according to this embodiment may improve the visibility of the contrast agent and medical device. However, bone image B may have more noise than stored image A, which may result in a problem of reduced image quality. Therefore, the radiation imaging system according to this embodiment performs noise reduction processing on bone image B.

[0075] 10 shows a block diagram of signal processing in the radiation imaging system according to the first embodiment. First, the signal processing unit 133 performs material separation processing in block MD1. Here, the material separation processing is processing in which the signal processing unit 133 acquires (generates) a bone image B and a soft tissue image S from a high-energy image H and a low-energy image L in the same procedure as in FIG. 6.

[0076] In block C2, the signal processing unit 133 generates a thickness image T by adding the bone image B and the soft tissue image S. Then, in block F1, the signal processing unit 133 performs a filtering process on the thickness image T to generate a noise-reduced thickness image T'. Details of the filtering process will be described in FIG. 11 and subsequent figures.

[0077] In addition, in block C1, the signal processing unit 133 generates an accumulated image A, for example, by multiplying an image H of the attenuation rate at high energy and an image L of the attenuation rate at low energy by a coefficient and adding them together, as described in the explanation of Figure 8.

[0078] Next, the signal processing unit 133 performs material separation processing again in block MD2. Hereinafter, the material separation processing again in block MD2 will also be referred to as re-separation (or re-separation processing). Herein, the re-separation processing is processing in which the signal processing unit 133 acquires (generates) a noise-reduced bone image B' from, for example, the thickness image T' after the filter processing and the stored image A.

[0079] The image of the sum of low-energy X-rays and high-energy X-rays, i.e., the spectrum of X-rays in accumulated image A, is expressed as N. A (E), if the thickness of the soft tissue is S and the thickness of the bone is B, the following equation

[11] holds.

[0080]

number

[0081] Here, if the thickness is T, then T=B+S, and by transforming equation

[11] , the following equation

[12] holds.

[0082]

number

[0083] By substituting the pixel value A and thickness T of the stored image at a pixel into equation

[12] and solving the nonlinear equation, it is possible to obtain the bone thickness B at a pixel. Substituting the filtered thickness T' instead of thickness T and solving equation

[12] , the bone thickness B' can be obtained. Because the thickness image T has higher continuity than the stored image A, it does not contain high-frequency components. Therefore, even if noise is removed by filtering, signal components are not easily lost. By using the noise-reduced thickness image T' and the originally low-noise stored image A, a noise-reduced bone image B' can be obtained. Similarly, a noise-reduced soft-tissue image S' can also be obtained.

[0084] In this embodiment, in the signal processing block diagram shown in Fig. 10, a configuration has been described in which a thickness image T is input to block F1 and a noise-reduced bone image B' or a noise-reduced soft-tissue image S' is acquired from block MD2. However, this embodiment is not limited to this example, and can be similarly applied to, for example, a configuration in which a soft-tissue image S is input to block F1 and a noise-reduced bone image B' or a noise-reduced thickness image T' is acquired from block MD2. Furthermore, this embodiment can be similarly applied to a configuration in which a bone image B is input to block F1 and a noise-reduced soft-tissue image S' or a noise-reduced thickness image T' is acquired from block MD2.

[0085] That is, in this embodiment, a material thickness image is input to block F1, and a material thickness image after noise reduction is obtained from block MD2.

[0086] In addition, in this embodiment, a configuration has been described in which a low-energy image L and a high-energy image H are input to block MD1 and block C1 in the signal processing block diagram shown in Fig. 10. However, this embodiment is not limited to this example, and can also be similarly applied to a configuration in which, for example, noise reduction processing is performed on the low-energy image L and the high-energy image H, respectively, to generate a noise-reduced low-energy image L' and a noise-reduced high-energy image H', and then input them to block MD1 and block C1.

[0087] The filtering process in block F1 includes spatial filtering using a spatial filter and temporal filtering using a temporal filter. Spatial filtering may use spatial filters such as a Gaussian filter, a bilateral filter, or an epsilon filter. Temporal filtering may use recursive filters. The filters that can be used in filtering are merely illustrative, and the configuration of the embodiment is not limited to these examples.

[0088] FIG. 11 is a diagram illustrating an example of noise reduction processing using a Gaussian filter according to the first embodiment. In block F1 of the signal processing block diagram shown in FIG. 10, the signal processing unit 133 uses a Gaussian distribution function to calculate weights according to the distance from the center of the filter and generates a Gaussian filter kernel. Furthermore, the signal processing unit 133 normalizes the sum of all weights of the Gaussian filter kernel to equal 1. The signal processing unit 133 then applies the Gaussian filter kernel to the thickness image T to generate a noise-reduced thickness image T'. In this manner, high-frequency noise components contained in the thickness image T can be removed. However, using a Gaussian filter in block F1 may result in the structure of the subject being reduced (removed) along with the noise components. As a result, when a noise-reduced bone image B' is generated by material separation (re-separation) in block MD2, the structure of the soft tissue S may appear in the bone image B'.

[0089] FIG. 12 is a diagram illustrating an example of noise reduction processing using an epsilon filter according to the first embodiment. In block F1 of the signal processing block diagram shown in FIG. 10, the signal processing unit 133 calculates weights according to distances from the center using a Gaussian distribution function to generate a kernel of a Gaussian filter. The signal processing unit 133 also selects a pixel to which the filter is applied (a pixel of interest) and calculates a difference d between the pixel of interest and its surrounding pixels in the thickness image T. If the difference d exceeds a predetermined threshold ε, the signal processing unit 133 determines that an object structure exists and sets the weight of the edge determination kernel to 0. If the difference d is below the predetermined threshold ε, the signal processing unit 133 determines that an object structure does not exist and sets the weight of the edge determination kernel to 1.

[0090] Next, the signal processing unit 133 generates an epsilon filter kernel by multiplying the Gaussian filter kernel by the edge determination kernel for each filter element. Furthermore, the signal processing unit 133 normalizes the sum of all weights of the epsilon filter kernel so that it becomes 1. Then, the signal processing unit 133 applies the epsilon filter kernel to the thickness image T to generate a noise-reduced thickness image T'. In this way, it is possible to reduce (remove) high-frequency noise components contained in the thickness image T without reducing (removing) the structure of the subject. This makes it possible to suppress the appearance of soft tissue structures in the noise-reduced bone image B'.

[0091] 13 is a diagram illustrating an example of edge determination processing using a stored image A according to the first embodiment. The inventors of the present application have conducted research and found that, in the configuration of block F1 shown in FIG. 12, for example, when the noise contained in the thickness image T increases, the accuracy of edge determination may decrease and artifacts may occur. Therefore, a configuration for performing edge determination using a stored image A instead of the thickness image T will be described below.

[0092] In block F1 of the signal processing block diagram shown in FIG. 10, the signal processing unit 133 uses a Gaussian distribution function to calculate weights according to distance from the center and generates a kernel of a Gaussian filter. The signal processing unit 133 also selects a pixel to which the filter is applied (a pixel of interest) and calculates a difference d between the pixel of interest and its surrounding pixels in the stored image A. If the difference d exceeds a predetermined threshold ε, the signal processing unit 133 determines that an object structure exists and sets the weight of the edge determination kernel to 0. If the difference d is below the predetermined threshold ε, the signal processing unit 133 determines that an object structure does not exist and sets the weight of the edge determination kernel to 1.

[0093] Next, the signal processing unit 133 generates an epsilon filter kernel by multiplying the Gaussian filter kernel by the edge determination kernel for each filter element. Furthermore, the signal processing unit 133 normalizes the sum of all weights of the epsilon filter kernel so that it becomes 1. Then, the signal processing unit 133 applies the epsilon filter kernel generated using the accumulated image A to the thickness image T to generate a noise-reduced thickness image T'. Since the accumulated image A has less noise than the thickness image T, for example, even if the noise contained in the thickness image T becomes large, the accuracy of edge determination can be maintained. As a result, the occurrence of artifacts can be suppressed.

[0094] The threshold value ε for edge determination is preferably set to, for example, ε>4σ, where σ is the standard deviation of noise that may be contained in pixel values ​​of stored image A. If the noise that may be contained in pixel values ​​follows a Gaussian distribution, the probability that the noise will cause a value of 4σ or more is extremely low, at 1 / 15787.

[0095] The noise that may be contained in pixel values ​​may include, for example, electrical noise (system noise) components and quantum noise components. When system noise is dominant, the standard deviation σ of the noise may be a constant value independent of the pixel value. Therefore, a configuration using a constant value as the threshold ε is preferably used. On the other hand, when quantum noise is dominant, the standard deviation σ of the noise may be a value proportional to the square root of the pixel value. Therefore, a configuration using a value proportional to the square root of the pixel value as the threshold ε is preferably used. The configuration of this embodiment makes it possible to obtain a material-separated image with reduced noise.

[0096] (Second embodiment) In the second embodiment, a configuration for removing artifacts such as low-frequency noise and false contours contained in a noise-reduced image B' will be described. In the second embodiment, the block diagram of the signal processing described in Fig. 10 etc. is the same as that in the first embodiment, but the content of the filter processing in block F1 is different.

[0097] FIG. 14 is a diagram illustrating an example of noise reduction processing using frequency decomposition according to the second embodiment. The inventors of the present application have conducted research and found that low-frequency noise may occur in the image T′ after noise reduction with the configuration of block F1 shown in FIG. 13. While this low-frequency noise is not particularly noticeable in a still image, it may reduce visibility in a moving image. Therefore, in this embodiment, a configuration for reducing low-frequency noise using frequency decomposition will be described.

[0098] The signal processing block F1 in FIG. 14 is made up of sub-blocks P[1] to P[n].

[0099] First, the thickness image T[1] is input to the sub-block P[1], and the signal processing unit 133 downsamples it using the decimation filter FD[1] to generate a thickness image T[2] with reduced resolution. Furthermore, the thickness image T[2] is input to the sub-block P[2], and the signal processing unit 133 downsamples it using the decimation filter FD[2] to generate a thickness image T[3] with reduced resolution. By repeating this process up to the sub-block P[n], the signal processing unit 133 generates frequency-resolved thickness images T[1] to T[n+1].

[0100] Furthermore, in the sub-block P[1], the signal processing unit 133 upsamples the thickness image T[2] using an interpolation filter FI[1] to generate a restored thickness image I[1] by restoring the decomposed frequency components (resolution). The signal processing unit 133 also reduces noise in the thickness image T[1] using an epsilon filter FE[1] to generate a noise-reduced thickness image E[1]. The epsilon filter FE[1] may be, for example, the epsilon filter described in FIG. 12. The signal processing unit 133 also subtracts the restored thickness image I[1] from the noise-reduced thickness image E[1] to generate a Laplacian image D[1]. The signal processing unit 133 repeats the same processing as that performed in the sub-block P[1] for the sub-blocks P[2] to P[n], thereby generating frequency-resolved Laplacian images D[1] to D[n].

[0101] Thereafter, in sub-block P[n], the signal processing unit 133 upsamples the frequency-synthesized image S[n+1] using an interpolation filter FC[n] to generate a restored frequency-restored image C[n]. The signal processing unit 133 also generates the frequency-synthesized image S[n] by adding the restored frequency-restored image C[n] to the Laplacian image D[n]. By repeating the same processing as that in the sub-block P[n] from sub-block P[n-1] to sub-block P[1], the signal processing unit 133 generates frequency-synthesized images S[n] to S[1].

[0102] 14, it is assumed that a frequency-resolved thickness image T[n+1] is input as a frequency-composite image S[n+1] in a sub-block P[n]. It is also assumed that a thickness image T is input as a frequency-resolved thickness image T[1] in a sub-block P[1], and that a frequency-composite image S[1] is output as a noise-reduced thickness image T'.

[0103] With the above configuration, the signal processing unit 133 performs frequency decomposition on the input thickness image T, performs noise reduction processing for each frequency using epsilon filters (FE[1] to FE[n]), and then generates a frequency composite image by combining all the frequency-resolved frequencies. As a result, it is possible to acquire a thickness image T' (frequency composite image S[1]) in which high-frequency noise to low-frequency noise contained in the thickness image T has been reduced.

[0104] FIG. 15 is a diagram illustrating an example of noise reduction processing using frequency decomposition and a recursive filter according to the second embodiment. The inventors of the present application have conducted research and found that with the configuration of block F1 shown in FIG. 14, for example, patterns resembling tree rings may be visible in areas where the thickness of the subject in the noise-reduced image T' changes suddenly. Such artifacts are called false contours. While false contours are not particularly noticeable in still images, they may reduce visibility in moving images. Therefore, this embodiment describes a configuration that uses frequency decomposition and a recursive filter to reduce false contours simultaneously with low-frequency noise.

[0105] The signal processing block F1 in Fig. 15 is also composed of sub-blocks P[1] to P[n], similarly to Fig. 14. The process from generating frequency-resolved thickness images T[1] to T[n+1] to generating frequency-resolved Laplacian images D[1] to D[n] is the same as the signal processing in Fig. 14.

[0106] In the signal processing block F1 of Figure 15, in the sub-block P[1], the signal processing unit 133 inputs the Laplacian image D[1][t] of the tth frame (t is an integer greater than or equal to 1) and the Laplacian image R[1][t-1] after recursive filter processing of the t-1th frame to the recursive filter FR[1] to obtain the Laplacian image R[1] after recursive filter processing of the tth frame (after temporal filter processing).

[0107] By repeating the same processing as that for sub-block P[1] for sub-blocks P[2] to P[n], the signal processing unit 133 generates Laplacian images R[1] to R[n] after recursive filtering (after time filtering).

[0108] Thereafter, in sub-block P[n], the signal processing unit 133 upsamples the frequency-synthesized image S[n+1] using an interpolation filter FC[n] to generate a restored frequency-synthesized image C[n]. The signal processing unit 133 also generates the frequency-synthesized image S[n] by adding the restored frequency-synthesized image C[n] to the recursively filtered Laplacian image R[n]. The signal processing unit 133 generates the frequency-synthesized images S[n] to S[1] by repeating the same processing as that for the sub-block P[n] from sub-block P[n-1] to sub-block P[1].

[0109] 14, in Fig. 15, the frequency-resolved thickness image T[n+1] is input as the frequency-composite image S[n+1] in the sub-block P[n]. Also, the thickness image T is input as the frequency-resolved thickness image T[1] in the sub-block P[1], and the frequency-composite image S[1] is output as the noise-reduced thickness image T'.

[0110] With the above configuration, the signal processing unit 133 frequency-decomposes the input thickness image T, performs noise reduction processing using epsilon filters (FE[1] to FE[n]) for each frequency, and further performs false contour reduction processing using recursive filters (FR[1] to FR[n]), and then generates a frequency-composite image by combining all the frequency-decomposed frequencies. As a result, it is possible to reduce high-frequency noise to low-frequency noise contained in the thickness image T and obtain a thickness image T' (frequency-composite image S[1]) with further reduced false contours. Then, the signal processing unit 133 generates an image by performing material separation processing again using the noise-reduced thickness image T' and the stored image A. With the configuration of this embodiment, it is possible to obtain a material-separated image with reduced noise.

[0111] (Third embodiment) In the third embodiment, a configuration will be described in which edge determination using stored image A is applied in the noise reduction processing using the epsilon filter in the second embodiment. In the third embodiment, the block diagram of the signal processing described in FIG. 10 etc. is the same as in the first embodiment. Also, the configuration of the frequency decomposition and recursive filter block diagram in FIG. 15 is the same as in the second embodiment. However, the configuration of the epsilon filters FE[1] to FE[n] is different.

[0112] Fig. 16 is a diagram illustrating an example of a configuration in which edge determination using stored image A is applied in noise reduction processing using an epsilon filter according to the third embodiment. In Fig. 15, the epsilon filter FE is configured to use thickness image T as described in Fig. 12. However, as described in Fig. 13, there may be cases in which a configuration in which edge determination processing using stored image A is more preferable.

[0113] Therefore, in this embodiment, the signal processing unit 133 generates a noise-reduced thickness image by filtering the thickness image of the material obtained by the material separation process using a filter whose coefficients are set based on the pixel values ​​of the accumulated image A obtained by adding together multiple images. The signal processing unit 133 downsamples the accumulated image A[1] using a decimation filter FA[1] to generate an accumulated image A[2] that has been frequency-resolved and has reduced resolution. Furthermore, the signal processing unit 133 downsamples the accumulated image A[2] using a decimation filter FA[2] to generate an accumulated image A[3] that has been frequency-resolved and has reduced resolution. By repeating this process, the signal processing unit 133 generates frequency-resolved accumulated images A[1] to T[n]. Note that in FIG. 16, the accumulated image A is input as the frequency-resolved accumulated image A[1].

[0114] In the sub-block P[1], the signal processing unit 133 inputs the frequency-resolved accumulated image A[1] and the frequency-resolved thickness image T[1] to the epsilon filter FE[1] to generate a noise-reduced thickness image E[1]. The epsilon filter FE[1] in the third embodiment is the epsilon filter shown in FIG. 13, which is generated by performing edge determination using the accumulated image. The signal processing unit 133 generates noise-reduced thickness images E[1] to E[n] by repeating the same processing as that in the sub-block P[1] for the sub-blocks P[2] to P[n]. Other processing is the same as that in the second embodiment described above.

[0115] With the above configuration, the signal processing unit 133 performs frequency decomposition on the input thickness image T, performs noise reduction processing using an epsilon filter for each frequency, and further performs false contour reduction processing using a recursive filter, and then generates a frequency composite image by combining all of the frequency-resolved frequencies. By frequency decomposing the accumulated image A and inputting it to the epsilon filter, the accuracy of edge determination is improved, and a thickness image T' with fewer artifacts can be obtained. Then, the signal processing unit 133 generates an image by performing material separation processing again using the thickness image T' after noise reduction and the accumulated image A. The configuration of this embodiment makes it possible to obtain a material-separated image with reduced noise.

[0116] In the first to third embodiments, the X-ray imaging device 104 is an indirect X-ray sensor using a phosphor. However, the disclosed technology is not limited to this. For example, a direct X-ray sensor using a direct conversion material such as CdTe may be used.

[0117] Furthermore, in the first to third embodiments, the tube voltage of the X-ray generator 101 is changed. However, the disclosed technology is not limited to this. The energy of the X-rays irradiated to the X-ray imaging device 104 may be changed by, for example, switching the filter of the X-ray generator 101 over time.

[0118] Furthermore, in the first to third embodiments, images of different energies were obtained by changing the energy of the X-rays. However, the disclosed technology is not limited to this configuration. A stacked type in which multiple phosphors 105 and two-dimensional detectors 106 are stacked may be used, thereby obtaining images of different energies from two-dimensional detectors on the front and back sides relative to the direction of incidence of the X-rays. The two-dimensional detector 106 is not limited to a medical detector, and may be an industrial two-dimensional detector.

[0119] In the first to third embodiments, the energy subtraction process is performed using the imaging control device 103 of the radiation imaging system. However, the disclosed technology is not limited to this configuration. The image acquired by the imaging control device 103 may be transferred to another computer, where the energy subtraction process is performed. For example, a configuration is preferably used in which the acquired image is transferred to another personal computer (image viewer) via a medical PACS, where the image is subjected to energy subtraction process and then displayed.

[0120] The disclosure of this specification includes the following image processing device, radiation imaging system, image processing method, and program.

[0121] (Item 1) A processing means is provided for performing noise reduction processing and material separation processing using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking an image, The processing means generating a noise-reduced image by applying a filter process to each frequency component obtained by decomposing the thickness image of the material obtained by the material separation process into a plurality of frequencies; An image processing device that generates an image by performing a second material separation process using a noise-reduced thickness image obtained by combining the noise-reduced images for each frequency component and an accumulated image obtained by adding up the multiple images.

[0122] (Item 2) The filtering process includes spatial filtering applied to each of the frequency components, 2. The image processing device according to item 1, wherein the processing means performs the spatial filtering process using an epsilon filter or a bilateral filter.

[0123] (Item 3) The image processing device according to item 1 or 2, wherein the processing means sets coefficients of a filter to be applied to the filtering process using the stored image decomposed into the frequency components.

[0124] (Item 4) The filtering process includes a time filtering process applied to each of the frequency components after the spatial filtering process, 3. The image processing device according to item 2, wherein the processing means performs the time filtering process using a recursive filter.

[0125] (Item 5) The image processing device according to Item 4, wherein the processing means performs the temporal filtering process on an image generated by subtracting a thickness image in which the frequency components are restored from a noise-reduced image to which the spatial filtering process has been applied.

[0126] (Item 6) The image processing device according to Item 5, wherein the processing means generates the noise-reduced thickness image by synthesizing a frequency composite image obtained by adding an image after time filtering processing and a frequency restored image obtained by restoring the frequency components in the thickness image for each of the decomposed frequency components.

[0127] (Item 7) The processing means applies the filter processing to a plurality of images corresponding to the plurality of different radiation energies to obtain the plurality of images after noise reduction, generating noise-reduced images by applying the filter processing to each frequency component obtained by decomposing a material thickness image obtained by the material separation processing using the plurality of images after noise reduction into a plurality of frequencies; 7. The image processing device according to any one of items 1 to 6, wherein an image in which substances contained in the subject are separated is generated by the substance separation process using a noise-reduced thickness image obtained by combining the noise-reduced images and an accumulated image obtained by adding the plurality of images.

[0128] (Item 8) A processing means is provided for performing noise reduction processing and material separation processing using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking an image, The processing means generating a noise-reduced thickness image by filtering the substance thickness image obtained by the substance separation process using a filter having a coefficient set based on pixel values ​​of the accumulated image obtained by adding the plurality of images; an image processing device that generates an image by performing a second material separation process using the noise-reduced thickness image and the stored image;

[0129] (Item 9) The processing means a weight for determining an edge of the structure of the subject based on whether a difference in pixel value between a pixel of interest and surrounding pixels in the stored image exceeds a threshold; 9. The image processing device according to item 3 or 8, wherein the coefficient of the filter is set based on the weight of the edge determination.

[0130] (Item 10) The image processing device according to Item 9, wherein the processing means sets the weight of the edge determination using a threshold that is independent of the pixel values ​​in the stored image or a threshold that is proportional to the square root of the pixel values ​​in the stored image.

[0131] (Item 11) The image processing device according to any one of items 1 to 10, wherein the processing means generates the stored image by multiplying at least one of the plurality of images by a coefficient and adding the multiplied images.

[0132] (Item 12) The processing means acquires the thickness image by adding a thickness image of a first substance obtained by the substance separation process and a thickness image of a second substance different from the first substance, 12. The image processing device according to any one of items 1 to 11, wherein a thickness image of the first substance is generated as an image obtained by the second substance separation process.

[0133] (Item 13) The processing means acquires a thickness image of a first substance obtained by the substance separation process and a thickness image of a second substance different from the first substance, inputting a thickness image of the second material as the thickness image; 12. The image processing device according to any one of items 1 to 11, wherein a noise-reduced thickness image is generated as the image obtained by the second material separation process by adding together the thickness image of the first material and the thickness image of the second material.

[0134] (Item 14) The image processing device according to item 12 or 13, wherein the thickness image of the first material is a bone image, and the thickness image of the second material is a soft tissue image.

[0135] (Item 15) The plurality of radiation energies include a first radiation energy and a second radiation energy lower than the first radiation energy, 15. An image processing device according to any one of items 1 to 14, wherein the processing means performs a material separation process using a first image captured with the first radiation energy and a second image captured with the second radiation energy to generate a first material separation image showing the thickness of a first material contained in the subject and a second material separation image showing the thickness of a second material different from the first material.

[0136] (Item 16) An image processing device that includes a processing means for generating a thickness image of a substance using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking an image, applying filter processing to each frequency component in the thickness image of the substance to generate a noise-reduced image for each frequency component, and synthesizing the noise-reduced images to generate a noise-reduced thickness image.

[0137] (Item 17) The image processing device according to any one of items 1 to 16, a radiation imaging device communicably connected to the image processing device;

[0138] (Item 18) A processing step of performing noise reduction processing and material separation processing using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking an image, In the processing step, generating a noise-reduced image by applying a filter process to each frequency component obtained by decomposing the thickness image of the material obtained by the material separation process into a plurality of frequencies; An image processing method for generating an image by performing a second material separation process using a noise-reduced thickness image obtained by synthesizing the noise-reduced images for each frequency component and an accumulated image obtained by adding up the multiple images.

[0139] (Item 19) A processing step is provided in which noise reduction processing and material separation processing are performed using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking an image, In the processing step, generating a noise-reduced thickness image by filtering the substance thickness image obtained by the substance separation process using a filter having a coefficient set based on pixel values ​​of the accumulated image obtained by adding the plurality of images; An image processing method for generating an image by performing a second material separation process using the noise-reduced thickness image and the accumulated image.

[0140] (Item 20) Image processing comprising a processing step of generating a thickness image of a substance using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking an image, applying a filter process to each frequency component in the thickness image of the substance to generate a noise-reduced image for each frequency component, and synthesizing the noise-reduced images to generate a thickness image after noise reduction.

[0141] (Item 21) A program that causes a computer to function as a processing unit of the image processing device according to any one of Items 1 to 16.

[0142] [Other embodiments] The disclosed technology can also be realized by supplying a program that realizes one or more functions of the above-described embodiments to a system or device via a network or a storage medium, and having one or more processors in the computer of the system or device read and execute the program. It can also be realized by a circuit (e.g., ASIC) that realizes one or more functions.

[0143] The disclosed technology is not limited to the above-described embodiments, and various changes and modifications can be made without departing from the spirit and scope of the invention. Accordingly, the following claims are appended to apprise the public of the scope of the invention. [Explanation of symbols]

[0144] 101: X-ray generator, 102: X-ray control device, 103: imaging control device (image processing device), 104: X-ray imaging device (radiation imaging device)

Claims

1. A processing means for performing noise reduction processing and substance separation processing using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating an object with radiation and performing imaging, The processing means is, Decompose the thickness image of the substance obtained by the substance separation process into a plurality of frequencies, Generate a noise-reduced image by applying a filter process for each of the decomposed frequency components, An image processing apparatus that generates an image by performing a substance separation process again using the noise-reduced thickness image obtained by synthesizing the noise-reduced image for each frequency component and the accumulated image obtained based on the synthesis of the plurality of images.

2. The filter process includes a spatial filter process applied for each of the frequency components, The image processing apparatus according to claim 1, wherein the processing means performs the spatial filter process using an epsilon filter or a bilateral filter.

3. The image processing apparatus according to claim 1, wherein the processing means sets a coefficient of a filter applied to the filter process using the accumulated image decomposed into the frequency components.

4. The filter process includes a temporal filter process applied for each of the frequency components after the spatial filter process, The image processing apparatus according to claim 2, wherein the processing means performs the temporal filter process using a recursive filter.

5. The image processing apparatus according to claim 4, wherein the processing means performs the temporal filter process on an image generated by subtracting a thickness image obtained by restoring the frequency components from the noise-reduced image by applying the spatial filter process.

6. The image processing apparatus according to claim 5, wherein the processing means generates the noise-reduced thickness image by synthesizing, for each of the decomposed frequency components, a frequency-synthesized image obtained by synthesizing an image after the temporal filter process and a frequency-restored image obtained by restoring the frequency components in the thickness image.

7. The processing means obtains the plurality of noise-reduced images by applying the filter process to the plurality of images corresponding to the plurality of different radiation energies, Decompose the thickness image of the substance obtained by the substance separation process using the plurality of noise-reduced images into a plurality of frequencies, Generate a noise-reduced image by applying the filter process for each of the decomposed frequency components, An image processing apparatus according to any one of claims 1 to 6, which generates an image in which substances included in the subject are separated by the substance separation process using the thickness image after noise reduction obtained by synthesizing the noise reduction images and the accumulation image obtained based on the synthesis of the plurality of images.

8. Comprising processing means for performing noise reduction processing and substance separation processing using a plurality of images corresponding to different radiation energies obtained by irradiating a subject with radiation and performing imaging, The processing means is, For the thickness image of the substance obtained by the substance separation process, a noise-reduced thickness image is generated by filter processing using a filter in which a coefficient is set based on the pixel value of the accumulation image obtained based on the synthesis of the plurality of images, An image processing apparatus that generates an image by performing a substance separation process again using the noise-reduced thickness image and the accumulation image.

9. The processing means is, Sets the weight of edge determination of the structure of the subject depending on whether the difference in pixel values between the pixel of interest and the surrounding pixels in the accumulation image exceeds a threshold value, The image processing apparatus according to claim 3 or 8, wherein the coefficient of the filter is set based on the weight of the edge determination.

10. The image processing apparatus according to claim 9, wherein the processing means sets the weight of the edge determination using a threshold value that does not depend on the pixel value in the accumulation image or a threshold value proportional to the square root of the pixel value in the accumulation image.

11. The image processing apparatus according to claim 1 or 8, wherein the processing means generates the accumulation image by multiplying at least one of the plurality of images by a coefficient and synthesizing them.

12. The processing means obtains the thickness image by the sum of the thickness image of the first substance obtained by the substance separation process and the thickness image of the second substance different from the first substance, The image processing apparatus according to claim 1 or 8, wherein the thickness image of the first substance is generated as the image by the substance separation process again.

13. The processing means obtains the thickness image of the first substance obtained by the substance separation process and the thickness image of the second substance different from the first substance, Inputs the thickness image of the second substance as the thickness image, The image processing apparatus according to claim 1 or 8, wherein a thickness image obtained by the sum of the thickness image of the first substance and the thickness image of the second substance, which is noise-reduced, is generated as the image by the substance separation process again.

14. The image processing apparatus according to claim 12, wherein the thickness image of the first substance is a bone image, and the thickness image of the second substance is a soft tissue image.

15. The plurality of radiation energies include a first radiation energy and a second radiation energy lower than the first radiation energy. The processing means performs a substance separation process of generating a first substance separation image indicating the thickness of a first substance included in the subject and a second substance separation image indicating the thickness of a second substance different from the first substance, using a first image taken with the first radiation energy and a second image taken with the second radiation energy. The image processing apparatus according to claim 1.

16. An image processing apparatus comprising processing means for generating a thickness image of a substance using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking a picture, applying a filter process for each frequency component in the thickness image of the substance to generate a noise reduction image for each frequency component, and generating a thickness image after noise reduction by synthesizing the noise reduction images.

17. An image processing apparatus according to any one of claims 1, 8, and 16, A radiation imaging apparatus communicably connected to the image processing apparatus, A radiation imaging system comprising.

18. A processing step of performing noise reduction processing and substance separation processing using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking a picture, In the processing step, Decompose the thickness image of the substance obtained by the substance separation process into a plurality of frequencies, Generate a noise reduction image by applying a filter process for each of the decomposed frequency components, An image processing method of generating an image by a second substance separation process using a thickness image after noise reduction obtained by synthesizing the noise reduction images for each frequency component and an accumulation image obtained based on the synthesis of the plurality of images.

19. A processing step of performing noise reduction processing and substance separation processing using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and taking a picture, In the processing step, A thickness image after noise reduction is generated by a filter process using a filter in which a coefficient is set based on the pixel value of an accumulation image obtained based on the synthesis of the plurality of images, for the thickness image of the substance obtained by the substance separation process. An image processing method for generating an image by performing a substance separation process again using the thickness image after noise reduction and the accumulation image.

20. Using a plurality of images corresponding to a plurality of different radiation energies obtained by irradiating a subject with radiation and performing imaging, a thickness image of a substance is generated, and by applying a filtering process for each frequency component in the thickness image of the substance, a noise reduction image for each frequency component is generated, and a process step of generating a thickness image after noise reduction by synthesizing the noise reduction images is provided. An image processing method.

21. A program that causes a computer to function as the processing means of the image processing apparatus according to any one of Claims 1, 8, and 16.

Images