JPH11149549A - Picture processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a picture processor for operating quick distortion correction. SOLUTION: A distortion correcting part 203 of a picture processor 202 generates a picture after correction from a picture before correction including geometrical distortion generated due to the characteristics of an X-ray photographic system for correcting distortion generated due to the characteristics of an image intensifier. Therefore, a picture after correction can be generated by the following procedures, that is, (1) the measurement of rough distortion distribution (measuring means), (2) the calculation of fine distortion distribution (calculating means), and (3) the decision of a concentration value (concentration value deciding means).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for correcting image distortion, and in particular, to various modalities such as an X-ray imaging apparatus or a computer tomography apparatus using an image intensifier (II). The present invention relates to an image processing apparatus suitable for correcting distortion of an image output from a computer.

[0002]

2. Description of the Related Art At present, X widely used in medical practice
Among the X-ray diagnostic apparatuses, an image intensifier is used as an X-ray detection system, which detects X-rays transmitted through a subject and converts the X-rays into an optical image.
A camera that shoots images using a V camera system is known. There are digital radiography and fluorography as X-ray imaging methods for digitizing image signals collected by this apparatus. A video signal from a TV camera is converted into a digital signal by an analog / digital (A / D) converter, and is input to a computer for performing various image processing.

An image intensifier has a structure in which an X-ray phosphor screen and a photocathode are superposed. The X-rays transmitted through the subject are input to the X-ray fluorescent screen, and light from the X-ray fluorescent screen stimulates the photocathode to emit photoelectrons, and a visible image (optical image) is formed on the output fluorescent screen. Imaged.

[0004] Such an image intensifier has a problem that its output image includes geometric distortion such as pincushion distortion and S-shaped distortion. For example, when a target object (phantom) as shown in FIG. 12A in which wires are arranged in a square lattice on a plane is photographed, FIG.
A distorted image is output as shown in FIG.

[0005] For example, when performing image processing for extracting useful information for diagnosis from an output image in the digital radiography, such distortion may be a major factor of an error. In particular, in clinical analysis such as calculation of a vascular stenosis rate and calculation of a left ventricle volume, or image processing such as three-dimensional reconstruction, the influence of an error due to distortion becomes remarkable.

[0006] In view of such a background, for example, a clinical analysis (see "Toshiba Digital Fluorography System"
CLINICAL ANALYSIS ", Toshiba Corporation, Toshiba Medical Corporation") or image processing such as three-dimensional reconstruction is appropriate and accurate, and image distortion caused by the image intensifier is indispensable. Conventionally, distortion correction is performed in the pre-processing stage of the image processing.

However, the conventional distortion correction has a problem that the processing takes an enormous amount of time because a fine distortion distribution is calculated for each pixel to obtain a corrected image. There is an urgent need for faster distortion correction in image processing in order to quickly provide high-precision diagnostic images in clinical settings and improve diagnostic performance.

[0008]

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide an image processing apparatus capable of performing distortion correction at high speed.

[0009]

(1) An image processing apparatus according to the present invention takes an image of a phantom having a specific geometric structure by an imaging system, and generates a geometric image generated due to the characteristics of the imaging system. An image processing apparatus for creating a corrected image from an uncorrected image including a large distortion, wherein a rough distortion distribution indicating a distortion point at a specific point specified according to a geometric structure of the phantom is captured by the phantom. Measuring means for measuring from an image, and calculating means for dividing the corrected image into a plurality of minute regions and calculating a fine distortion distribution indicating a distortion point of a representative point in the minute region using the coarse distortion distribution And density value determining means for determining the density value of the minute area on the image after correction based on the density value of the minute area on the image before correction specified by the distortion point of the representative point. Have.

According to such a configuration, the position of the pixel is associated with each minute region between the image before correction and the image after correction in a fine distortion distribution, so that the position is associated with each pixel. Thus, the calculation time can be greatly reduced. Therefore, the speed of the distortion correction process can be increased. This is based on the assumption that there is almost no distortion in the minute area. (2) The image processing apparatus according to the present invention is the apparatus according to the above (1), wherein the calculating means includes a coarse distortion indicating distortion destinations of at least four specific points near a representative point in the minute area. It is characterized in that a fine distortion distribution indicating a distortion point of the representative point is calculated by performing a two-dimensional interpolation calculation on the distribution. (3) The image processing apparatus according to the present invention is the apparatus according to the above (1), wherein the calculating unit performs the correction after correcting the plurality of minute areas having different sizes at the center and the periphery of the image. And a means for calculating a fine distortion distribution indicating a distortion point of a representative point in each of the minute regions based on the coarse distortion distribution. (4) The image processing apparatus according to the present invention is the apparatus according to the above (1), and further includes a setting unit that sets a region of interest on the corrected image, and the calculation unit performs the setting. Means for dividing the image after correction into a plurality of minute regions having different sizes inside and outside the region corresponding to the region of interest set by the means, and based on the coarse strain distribution, Means for calculating a fine strain distribution indicating the strain point of the representative point in the region. (5) The image processing apparatus according to the present invention is the apparatus according to the above (1), wherein the density value determining means calculates the corrected image according to a distortion point of the representative point calculated by the calculating means. Means for specifying a minute area on the image before correction corresponding to the above minute area, based on the density value in the minute area on the image before correction specified by this means,
Means for determining a density value in a minute area on the image after the correction. (6) The image processing apparatus according to the present invention is the apparatus according to the above (1), wherein the phantom has a lattice-like structure, and the measuring unit is configured to control the phantom on the corrected image. The distance between the grid points is converted to an integer, and a coarse strain distribution indicating the distortion destination of each of the converted grid points is measured.The calculating means determines that the representative point of the plurality of minute regions is located at the center of the pixel. In such a case, the image after the correction is divided. (7) The image processing apparatus according to the present invention is the apparatus according to the above (6), and further includes means for enlarging or reducing the corrected image by changing a distance between the grid points. It is characterized by. (8) The image processing apparatus according to the present invention is the apparatus according to the above (7), wherein the distance between the lattices is converted to the nearest integer.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the image processing apparatus according to the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram schematically showing a configuration of an X-ray diagnostic system according to an embodiment of the image processing apparatus of the present invention. The system is X
This is a system for performing image processing including distortion correction on an X-ray contrast image of a subject obtained from a radiographic system, and providing the obtained image to a diagnosis by a doctor or the like.

[0012] As shown in FIG.
X-ray source device 1 including ray tube 100 and irradiation field limiter 101
02, an image intensifier that detects an X-ray beam emitted from the X-ray tube 100 to the subject M via the irradiation field limiter 101 and transmitted through the subject M, and converts the X-ray beam into an optical image ( I.I.) 103 and I.I. I. An X-ray imaging system includes an X-ray TV camera 104 that captures an optical image output from the digital camera 103 and outputs a video signal corresponding to the optical image. The video signal output from the TV camera 104 is converted into a digital video signal by the A / D converter 201 and stored in the image memory 206 including the image storage unit 204 and the auxiliary information storage unit 205.

Image data is stored in the image storage unit 204. Patient information such as the patient's name, gender, date of birth, and weight, and various information attached to the image data, ie, II- Additional information indicating imaging conditions such as an X-ray focal length, an enlargement ratio, an II size, an imaging position, and an imaging angle is stored.

The image processing device 202 includes an image memory 206
The X-ray image of the subject read from the
Various types of image processing including distortion correction due to
The image immediately after being output from the A / D converter 201 is I.D.
I. The distortion caused by the characteristic of the reference numeral 103 is included. The distortion correction unit 203 of the image processing device 202 reads the image data before correction from the image storage unit 205 of the image memory 206, processes the image data according to a processing process described later, and corrects the distortion corrected by the processing. Create post-image.

The created image data after the distortion correction is output to the image display unit 207. The image display unit 207 outputs the image data after the distortion correction to the monitor 210 or the laser imager 211. Thus, the X of the subject after distortion correction
A line diagnostic image is provided for display.

A computer (Central Processing)
Unit: not shown) controls the operation of the A / D converter 201, the image processing device 202, and the image display unit 207. Further, the image data of the image after the distortion correction can be recorded on a magnetic disk (or optical disk) 209 having a large capacity according to an instruction from a computer. It should be noted that the additional information stored in the additional information storage unit 205 can be recorded together with the image recording.

I. I. In order to correct the distortion caused by the characteristic of the image processing unit 103, the distortion correction unit 20 of the image processing apparatus 202
3 executes a distortion correction process as follows. That is, the distortion correction unit 203 creates a corrected image in which the distortion has been corrected from the image before correction including the geometric distortion generated due to the characteristics of the X-ray imaging system according to the following procedure. .

(1) Measurement of coarse strain distribution (measurement means) (2) Calculation of fine strain distribution (calculation means) (3) Determination of density value (density value determination means) A coarse strain distribution indicating a distortion point at a specific point specified according to the geometric structure of the phantom is measured from the captured image of the phantom on the subsequent image. The calculating means of (2) divides the corrected image into a plurality of minute areas, and calculates a fine strain distribution indicating a distortion point of a representative point in the minute area based on the coarse strain distribution measured in (1). And calculate. The density value determining means of (3) determines the density value of the minute area on the image after correction based on the density value of the minute area on the image before correction specified by the distortion point of the representative point. .

According to the distortion correction of the present embodiment configured as described above, in the fine distortion distribution, the position of the pixel is associated with each minute region between the image before correction and the image after correction. The calculation time can be significantly reduced as compared with the case where the position is associated with the position. Therefore, the speed of the distortion correction process can be increased. This is based on the assumption that there is almost no distortion in the minute area.

Here, an image before distortion correction (hereinafter referred to as a “distorted image”)
Image) and the image after correction (hereinafter referred to as "corrected image")
The various parameters relating to both are specifically defined, and the distortion correction processing will be described in detail.

First, image coordinates are defined as shown in FIG. The coordinates (i, j) represent a two-dimensional position of a pixel on the digital image input to the image processing device 202. The unit of i and j is [pixel].
For example, in an image having an image size of N × N, as shown in the figure, the upper left pixel is (1, 1), the upper right pixel is (N, 1), and the lower left pixel is (1, N). , The lower right pixel of the image is represented as (N, N). Practical I. I. Then
In many cases, the image size is set to “1024 × 1024”. Alternatively, there is a case where “512 × 512” or “2048 × 2048” is set.

Next, the density values of the distorted image and the corrected image are defined as follows. D (i, j): density value of pixel at coordinates (i, j) of distorted image C (i, j) density value of pixel at coordinates (i, j) of corrected image Uses a grid arranged in a grid pattern, and grid coordinates are defined as shown in FIG. FIG. 3A shows a grid projection image when there is no distortion, and FIG. 3B shows a grid projection image when there is distortion.

Next, a coarse strain distribution measured from the grid projection image is defined as follows. G￣ (u, v): Image coordinates (two-dimensional) of lattice point (u, v) on strain image G _i (u, v): Horizontal of image coordinates of lattice point (u, v) on strain image (I) Direction component G _j (u, v) ... horizontal (j) direction component of the image coordinates of the lattice point (u, v) on the distortion image, where (u, v) is a lattice point, ie, a wire This is a grid coordinate representing the position of the intersection. Here, the lattice point closest to the center of the distorted image is set to the origin (i ₀ ,
j ₀ ), and the origin of the grid coordinates is set to (u ₀ , v ₀ ) as shown in FIG. Note that “G な お” represents a vector.

For example, if the grid points are 5 × 5, FIG.
As shown in (b), the coordinates of the upper left grid point, that is, the grid coordinates are (1, 1), the grid coordinates of the upper right grid point are (5, 1), and the grid coordinates of the lower left grid point are (1, 1).
5), the grid coordinates of the lower right grid point are represented as (5, 5).

A certain grid point (u, v) exists at different positions (different image coordinates) on the distorted image and the corrected image, and the correspondence of the image coordinates between the two images represents the distortion distribution. In other words, the lattice point should be projected at the position of the lattice point (u, v) on the corrected image unless distortion originally occurs. I. , The position Gｕ (u, v) of the lattice point (u, v) on the distorted image
Is projected to

Next, a fine strain distribution calculated based on the coarse strain distribution is defined as follows. Coarse strain distribution G￣
The corresponding relationship between a predetermined sample point (representative point) on the corrected image calculated from (u, v) and a point (distortion point) corresponding to the representative point on the distorted image is P￣ (i, j). And defined as follows.

P￣ (i, j): Image coordinates of the distortion point of the sample point (i, j) on the corrected image (on the distortion image: two-dimensional) P _i (i, j): Sample points on the corrected image The horizontal (i) direction component of the image coordinates of the distortion destination (i, j) P _j (i, j) ... the vertical (j) direction component of the image coordinates of the distortion destination of the sample point (i, j) on the corrected image The predetermined sample point is a sample point in a predetermined grid pattern on the corrected image.

The measurement of the coarse strain distribution in the above (1) is specifically performed as follows. First, the grid is called I.D.
I. X-ray imaging is performed by sticking to the front surface of 103, and image coordinates of all grid points (intersections of wires) are obtained from the obtained images. Next, I. I. 1
Based on the enlargement ratio of 03, the image matrix size, and the distance between adjacent wires, the distance L between wires on the captured image (the unit is [pixel]: L may be a decimal number) is determined.

The measurement of the coarse strain distribution is performed in an environment such as geomagnetism,
As long as there is no change in the installation location, imaging angle, etc. of the X-ray imaging system, it is not necessary to measure each time during X-ray imaging (during image processing).
May be used.

The phantom having the geometric structure is not limited to the grid used here. For example, the measurement may be performed using a phantom in which two-dimensionally small spheres are arranged at equal intervals. For a description of this measurement method, see the reference “Beth Schueler et al.,” “Correction of
image intensifier distortion for 3D X-ray angiogra
phy ", Proceedings of SPIE's Medical Imaging, Vol.2
432 (Physics of Medical Imaging). Pp272-279 (1995) ”
Can be referred to.

The measurement of the fine strain distribution in the above (2) is specifically performed as follows. Here, based on the measured coarse distortion distribution, it is calculated which pixel on the distorted image corresponds to the pixel indicating the representative point in the plurality of minute regions in the corrected image.

The density value of C (i, j) is obtained by the following procedure. First, on the corrected image, FIG.
Let us consider a grid lattice pattern of L × L [pixels] from the image center (i ₀ , j ₀ ) as shown in FIG.

Next, sampling is performed at substantially equal intervals in the corrected image to obtain a plurality of sample points. At this time, the sample point is set at the pixel center position. FIG. 4 is a diagram for explaining such a sampling process. FIG. 4 (a)
Shows an overall view of a grid grid pattern and an arbitrarily selected grid pattern (within a dotted line), and FIG. 4B shows an enlarged view of the selected grid pattern. In FIG. 4B, black circles represent sample points, and a hatched area represents a minute area having the sample point as a representative point (representative pixel). Incidentally, since the distance between grid points is generally a decimal number, the grid boundary does not pass through the center of the pixel. In such a case, the pixel passing through the boundary is made to belong to the larger lattice area divided by the boundary. However, in such a case, since the shape (size) of the grid is slightly different depending on the location, it is necessary to change the method of dividing the micro area for each grid or to store the change result in a memory. Each of these causes an increase in processing time and memory occupation. For such a problem, if the distance between the grids is converted to the nearest integer, the shapes of the grids can all be made the same. Therefore,
All the grids can be similarly divided simply by storing the way of dividing one grid into minute regions. This allows
The calculation time can be shortened and the required memory capacity can be reduced, which is superior to the case where the distance between lattice points is set to a decimal number.

Here, the grid pattern is divided so that each pixel belongs to the closest sample point. A pixel surrounded by a dotted line is divided so as to be closest to a sample point at the center of the other sample points. In the example of FIG. 4, the data is divided as indicated by a dotted line. Pixels in one grid surrounded by the dotted line are divided so as to be closest to the sample point at the center of the other sample points.

The strain distribution at the sample point (i, j) is obtained by the following procedures 1 to 6. Step 1: Find the amount of deviation from the origin used for processing. In the example of FIG. 5, for example, any sample point (i, j) shift of the image coordinates from to the center _{(i-i 0, j-} j 0), is converted into displacement of a grid pattern.

Δu = (int) ｛(i−i ₀ ) / L｝ Δv = (int) ｛(j−j ₀ ) / L｝ (1) where the (int) operator has the following numerical value. If the value is positive, the value after the decimal point is rounded down and converted to an integer. If the value is negative, the value after the decimal point is rounded up and converted to an integer. For example, if the numerical value is -1.2, it becomes -2 after conversion. Step 2: Strain distribution data (4 points) used for the processing is obtained. In step 1, the shift amount (Δu, Δv) from the center of the image to the upper left grid point of the grid pattern including (i, j) was determined. Therefore, the grid coordinates of the upper left grid point can be expressed as (u ₁ , v ₁ ) by the following equation. Similarly, the grid coordinates of the upper right, lower left, and lower right points are (u ₂ , v ₁ ), (u ₁ , v ₂ ), and (u ₂ , v ₂ ).

U ₁ = u ₀ + Δuv ₁ = v ₀ + Δv u ₂ = u ₁ +1 v ₂ = v ₁ +1 (2) Step 3: Distortion point at point (i, j) Based on For this purpose, first, as shown in FIG. 6 (a), the internal division ratio of the point (i, j) in the grid pattern is determined in the horizontal and vertical directions of the image, and these are respectively calculated as m
_i : _ni , _mj : _nj .

[0038] _{m i = {i- (i 0} + Δu · L)} / L n i = 1-m i ... (3) m j = {j- (j 0 + Δv · L)} / L n j = 1 −m _j (4) FIG. 6B shows a grid corresponding to the grid of FIG. 6A, and grid (grid) coordinates of points P, Q, R, and S, and P, R, The image coordinates of Q, S, T, U, and V are as follows, respectively.

The grid (grid) coordinates of P (G _i (u ₁ , v ₁ ), G _j (u ₁ , v ₁ )) The grid (grid) coordinates of Q (G _i (u ₂ , v ₁ ), G _j (u ₂ , v ₁ )) R grid (grid) coordinates (G _i (u ₁ , v ₂ ), G _j (u ₁ , v ₂ )) S grid (grid) coordinates (G _i (u ₂ , v ₂ ), G _j (u ₂ , v ₂ )) P image coordinates (P _i (i, j ₁ ), P _j (i, j ₁ )) R image coordinates (P _i (i ₁ , j ₂ ), P _j (i ₁ , j ₂ )) Q image coordinates (P _i (i ₂ , j ₁ ), P _j (i ₂ , j ₁ )) S image coordinates (P _i (i ₂ , j ₂ ), P _j (i ₂ , j ₂ )) T image coordinates (P _i (i, j ₁ ), P _j (i, j ₁ )) U image coordinates (P _i (i, j ₂ )) , P _j (i, j ₂ )) V image coordinates (P _i (i, j), P _j (i, j)) where two-dimensional interpolation (first order Use interpolation function: Bi-Lin
By applying (ear interpolation), the distortion destination of the point (i, j) is obtained as follows. As shown in FIGS. 7A and 7B, the calculation is independently performed in the i direction and the j direction.

[1] From the distortion distribution G￣ (u ₁ , v ₁ ) measured from the grid projection image, G￣ (u ₂ , v ₁ ) to P￣
Estimate (i, j ₁ ). P _i (i, j ₁ ) = m _i G _i (u ₂ , v ₁ ) + n _i G _i (u ₁ , v ₁ ) P _j (i, j ₁ ) = m _i G _j (u ₂ , v ₁₎ ) + N _i G _j (u ₁ , v ₁ ) (5) [2] Similarly to the above [1], the measured strain distribution G￣ (u
₁ , v ₂ ), G￣ (u ₂ , v ₂ ) to P￣ (i, j ₂ )
Is estimated. P _i (i, j ₂ ) = m _i G _i (u ₂ , v ₂ ) + n _i G _i (u ₁ , v ₂ ) P _j (i, j ₂ ) = m _i G _j (u ₂ , v ₂₎ ) + N _i G _j (u ₁ , v ₂ ) (6) [3] Estimated strain distribution P￣ (i, j ₁ ), P￣ (i, j)
₂ ) Estimate the strain distribution P ， (i, j) that is ultimately desired to be obtained. P _i (i, j) = m _j P _i (i, j ₂ ) + n _j P _i (i, j ₁ ) P _j (i, j) = m _j P _j (i, j ₂ ) + n _j P _j (I, j ₁ )... (7) Step 4: Coordinates (P _i (i, j), P _j (i,
Since P _i (i, j) and P _j (i, j) in j)) are not integers in most cases, a sample point of the image closest to this point is obtained. _{i p = (int) (P} i (i, j)
_{+0.5), j p = (int} ) (P j (i, j) +0.5) ...: relative (8) Step 5 areas divided on the corrected image (see FIG. 8 (a) hatched portion) a region having the same size, set in the center Ibitsusaki coordinates (i _p, j _p) to the distorted image (FIG. 8
(A hatched portion in (b)), assuming that there is no distortion inside, the value on the distorted image is substituted into the value of the corrected image according to the following equation. Thereby, the density value of the corrected image is determined. C (i + Δi, j + Δj) = D (i p + Δi, j p + Δj) (-L / 2 ≦ Δi, Δj ≦ L / 2 or microscopic area) ... (9) Step 6: all covering the upper corrected image Steps 2 to 5 are repeated for the minute area.

The corrected image C (i, j) obtained by the distortion correction unit 203 as described above is stored in the image data storage unit 204. Here, when the distortion correction according to the present embodiment is compared with the first and second comparative examples using a method different from the distortion correction in terms of the amount of calculation, the calculation result of the entire image is as follows. Note that the number of pixels in the entire image is N, and the number of pixels in the minute area is M. The first comparative example in which the strain distribution is obtained by first-order approximation and the density is also obtained by first-order approximation Integerization 4N, addition 17N, subtraction 10N, multiplication 20N, division 4N Second comparative example obtained by zero-order approximation Integerization 4N, addition 14N, subtraction 6N, multiplication 14N, division 4N ・ Distortion correction of this embodiment Integerization 4N / M, addition 14N / M, subtraction 6N / M, multiplication 14N / M, division 4N / M If the minute area is 3 × 3 pixels, the amount of calculation required for distortion correction is 1/9 of the amount of calculation required in the second comparative example. Higher speed is possible.

Here, various modifications of the above embodiment will be described. (Modification 1) In the above embodiment, the entire image is sampled in the same manner. However, the amount of distortion is smaller at the center of the image, and becomes larger as it goes outside the image. Therefore, in the vicinity of the center of the image, the number of calculations can be further reduced by reducing the sampling, that is, by setting the minute area to be relatively large. In the above embodiment, the minute area is set to, for example, 3 × 3 pixels in all the grid patterns. However, in a region of 縦 of the image centered on the center of the image (1/4 of the entire image in area), the small region is set to 5 × 5 pixels assuming that the distortion amount is small, and the other regions are set to 3 × 5 pixels as usual. Assume 3 pixels. Then, the amount of calculation can be reduced by about 15% compared to the case of the above embodiment, and as a result, about 1.18 times the speed can be achieved.

(Modification 2) In the above embodiment, the entire image is sampled in the same manner. However, information important for diagnosis exists near the center of the image, and outside of the information only needs to display the relationship with the region of interest (which can be set as a region of interest using a ROI marker or the like). In such a case, some discontinuity in the structure (such as block distortion like DCT) does not cause much problem, and it suffices if the position can be specified from the approximate structure. Therefore, in the vicinity of the image, sampling is small,
That is, the amount of calculation can be further reduced by setting the micro area to be relatively large. In the above embodiment, the minute area is set to, for example, 3 × 3 pixels in all the grid patterns. However, this is outside the area of 縦 of the image centered on the image center (3/3 of the entire image in area).
In 4), a small area is assumed to be 5 × 5
Pixels, and the others are 3 × 3 pixels as usual. Then, the operation amount can be reduced by about 50% as compared with the above embodiment, and as a result, about twice as high speed can be achieved.

(Modification 3) In the above embodiment, it is assumed that the size of the minute region of the image before correction is the same as the size of the minute region of the image after correction. However, actually, the periphery of the image is enlarged compared to the center of the image due to the influence of the pincushion distortion. Here, the enlargement ratio at the center of the image is set to 1, and the enlargement ratio at an arbitrary position in the image is defined as the distance projected to one pixel at the center of the image (for example, the length of the wire in front of the detector). It is expressed by the number of pixels projected. In this case, the enlargement ratio is expressed as a function of the distance from the center of the image, for example, as shown in FIG.
Here, the horizontal axis is the distance from the center of the image, where 0% indicates the center of the image and 100% indicates the periphery of the image. In the case where the minute area is relatively small, for example, 2 × 2 [pixel ² ], the difference in the magnification does not affect the corrected image so much that the above embodiment can sufficiently cope with it. When the minute region is enlarged, the influence of the difference in the magnification increases, and discontinuity in the human body structure occurs at the boundary of the minute region. This can be solved by configuring the image processing apparatus as shown in FIG. The image processing device 300 shown in FIG. 10 includes an affine transformation unit 301, an image memory unit 302, and a distortion correction unit 303.

Here, for example, I. I. Has an enlargement characteristic as shown in FIG. 9 (in this example, the maximum enlargement ratio is 1.
6). First, the distorted image read from the image memory 206 is input to the image processing device 300.
The affine transformation unit 201 creates eight reduced images by changing the reduction ratio every 0.05 from 0.95 to 0.60 times, and stores these reduced images and the original image in the image memory unit 302.
To store temporarily.

Next, the distortion correction unit 303 performs processing substantially in the same manner as in the first embodiment. In particular, here, a minute area is set on a distortion image as follows. That is, a distorted image whose reciprocal of the enlargement ratio at the sampling (representative) point is closest to the reduction ratio is selected, and the minute area data is derived from the image. This means that the reduced (enlarged) uncorrected image has a size substantially the same as that of the minute region (not in the image but in the real space) so as to cancel the difference in the enlargement ratio of each minute region due to the influence of the pincushion distortion. Of the size).

(Modification 4) If the distance L between the wires is smaller than the actual value in Modification 3, the properties of the distortion correction method according to the present invention (exactly, the method of obtaining the value by zero-order approximation when obtaining the value) Due to the nature), there is a high probability that a pixel that exists on the distorted image but cannot contribute to the corrected image is generated. Therefore, the wire spacing L
Is desirably set larger than the actual value.

For example, consider expanding the wire interval L to 1.6 times the actual value based on the expansion characteristics as shown in FIG. That is, the affine transformation unit 301 creates 12 enlarged images by changing the enlargement ratio by 0.05 to 1.05 to 1.6 times every 0.05.
The data of the minute area is derived from the image magnified 6 times, and the data of the minute area is derived from the original image around the image.
The image area between the image center and the image periphery is described in I.C.
I. In consideration of both the enlargement due to the pincushion distortion and the image enlargement, the data of the minute area is derived from the image corresponding to the largest enlargement ratio.

The above-described enlargement of the corrected image is merely an example. According to the same processing, a corrected image having an arbitrary enlargement ratio can be created. The enlargement of the image is not limited to the realization method according to the present modification, and for example, the corrected image may be directly enlarged.

(Modification 5) In the above-described first embodiment, as shown in FIG. 4B, a minute area is set by dividing the grid grid at equal intervals. As shown in FIG. 11, it may be set so as to straddle between grid grids. In this case, since the distortion destination data of the lattice point can be used as it is as the fine distortion distribution, and the fine distortion distribution on the lattice boundary can be obtained by the first-order interpolation, the total number of the fine distortion distribution data is the case of the first embodiment. If this is not the case, it is possible to achieve a higher speed than in the first embodiment.

The present invention is not limited to the above embodiment,
Various modifications are possible. For example, in the above-described embodiment, an X-ray diagnostic system in which an X-ray contrast image of a subject is obtained by an X-ray imaging system has been described. However, other modalities different from the X-ray imaging system, such as I.D. I. X-ray computed tomography apparatus equipped with

[0052]

As described above, according to the present invention, it is possible to provide an image processing apparatus capable of performing distortion correction at high speed. According to such an image processing apparatus, it is possible to operate the apparatus in a clinical setting such as a normal examination routine. Therefore, the accuracy of processing results such as image analysis and measurement can be improved, and information for performing accurate diagnosis can be provided to a clinical setting.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing a configuration of an X-ray diagnostic system according to an embodiment of an image processing apparatus of the present invention.

FIG. 2 is a diagram showing a definition of image coordinates.

FIG. 3 is a diagram showing a grid projection image.

FIG. 4 is a view showing a lattice pattern.

FIG. 5 is a diagram showing positions in a strain distribution to be used.

FIG. 6 is a diagram showing a correspondence relationship between lattices.

FIG. 7 is a diagram showing calculation of a distortion destination.

FIG. 8 is a diagram showing a correspondence relationship between a corrected image and a distorted image.

FIG. 9 is a graph showing an expansion characteristic according to a third modification.

FIG. 10 is a block diagram illustrating a configuration of an image processing apparatus according to a third modification.

FIG. 11 is a diagram showing a method for dividing a minute area according to a fifth modification.

FIG. 12 is a diagram showing a mode of distortion according to a conventional example.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 ... Irradiation field limiter 102 ... X-ray source device 103 ... Image intensifier (II) 104 ... TV camera 201 ... A / D converter 202 ... Image processing device 203 ... Distortion correction part 204 ... Image storage unit 205 ... Attachment information storage unit 206 ... Image memory 207 ... Image display unit 208 ... Computer 209 ... Magnetic disk 210 ... Monitor 211 ... Laser imager

Claims (8)

[Claims] 1. A phantom having a specific geometric structure is photographed by a photographing system, and a corrected image is created from an uncorrected image including a geometric distortion caused by characteristics of the photographing system. In the image processing apparatus, a measuring means for measuring, from a captured image of the phantom, a coarse strain distribution indicating a distortion point of a specific point specified according to a geometric structure of the phantom; and a plurality of the corrected images. Calculating means for calculating a fine strain distribution indicating a distortion point of a representative point in the minute area using the coarse strain distribution, and a density value of the minute area on the image after the correction. And a density value determining unit for determining based on a density value of a minute area on the image before correction specified by a distortion point of the representative point. 2. The method according to claim 1, wherein the calculating unit performs a two-dimensional interpolation operation on a coarse distortion distribution indicating distortion points of at least four specific points near the representative point in the minute area, thereby obtaining a fine distortion point indicating the distortion point of the representative point. The image processing device according to claim 1, wherein a distortion distribution is calculated. 3. The method according to claim 2, wherein the calculating unit divides the corrected image into a plurality of minute regions having different sizes at a central portion and a peripheral portion of the image. 2. The image processing apparatus according to claim 1, further comprising: means for calculating a fine distortion distribution indicating a distortion point of a representative point in the minute area. 4. The image processing apparatus according to claim 1, further comprising: a setting unit configured to set a region of interest on the corrected image, wherein the calculating unit sets a size of the region inside and outside the region corresponding to the region of interest set by the setting unit. Means for dividing the corrected image into a plurality of different small areas, based on the coarse strain distribution, means for calculating a fine strain distribution indicating a distortion point of a representative point in each of the fine areas, The image processing apparatus according to claim 1, further comprising: 5. The density value determining means specifies a minute area on the image before correction corresponding to the minute area on the corrected image according to a distortion point of the representative point calculated by the calculating means. And a means for determining a density value in a minute area on the image after correction based on a density value in the minute area on the image before correction specified by the means. The image processing device according to claim 1. 6. The phantom has a grid-like structure, and the measuring means converts the distance between the grid points of the phantom on the corrected image into an integer, and calculates the individualized grid points. Measuring the coarse distortion distribution indicating the distortion destination of the image, wherein the calculating unit divides the corrected image so that a representative point of the plurality of minute regions is located at the center of a pixel. 2. The image processing device according to 1. 7. The image processing apparatus according to claim 6, further comprising: means for enlarging or reducing the corrected image by changing a distance between the grid points. 8. The image processing apparatus according to claim 7, wherein a distance between the grids is set to a nearest integer.