JP5051025B2 - Image generating apparatus, program, and image generating method - Google Patents

Description

  The present invention relates to an image generation technique.

  In the medical field, various examinations and diagnoses are performed by photographing an affected part included in a built-in structure or a skeleton using an X-ray or the like. In recent years, it has become possible to relatively easily acquire a moving image that captures the motion of an affected area using X-rays or the like by applying digital technology.

  An example of an organ that is effective for diagnosing the movement of the affected area is, for example, a lung whose shape changes greatly due to respiration. For example, the lungs tend to have a significantly reduced movement of expansion and contraction in areas with disease. For this reason, the doctor can make a diagnosis by recognizing the behavior of the lung through the moving image.

  Also, as a technique for analyzing the motion of the affected area, a difference between temporally adjacent X-ray images using a plurality of images (X-ray images) photographed using X-rays continuously in time series. A technique for acquiring (difference image) has been proposed (for example, Patent Document 1). In addition, a technique for removing a noise component from an X-ray image has been proposed (for example, Non-Patent Document 1).

JP 2004-31434 A "Evaluation of Pulmonary Function Using Breathing Chest Radiography With a Dynamic Flat Panel Detector", Rie Tanaka et al., Investigative Radiology, vol.41 (10) p735-745, October 2006.

  However, in the technique of the above-mentioned Patent Document 1, a temporal change in blood flow due to the pulsation of the heart is superimposed on a moving image capturing the lung, and the local brightness of the moving image capturing the lung is blood. Since it changes according to flow fluctuations, it becomes difficult to distinguish the features of the lung from other features.

  In the technique of Non-Patent Document 1, attention is paid to a change in a pixel value with respect to time for each pixel of interest in a moving image (chest X-ray moving image) obtained by imaging a chest using X-rays. By removing the high-frequency component in, the fluctuation component due to blood flow is being deleted.

  However, since the region of the subject (lung field region) fluctuates due to breathing, the possibility that the same lung field region appears in the same pixel in a plurality of frames of the chest X-ray moving image is low. For this reason, when the high-frequency component is simply removed from the same pixel, not only the fluctuation component due to blood flow but also other fluctuation components, that is, the fluctuation component of the lung due to respiration tend to be removed. In other words, it is difficult to visually distinguish the features of the lung from other features with high accuracy.

  Such a problem is not only in the case of acquiring an image that captures the lung using X-rays, but also under a condition in which a variation component due to an element other than the specific region is superimposed on the feature of the specific region using various techniques. It is generally common when acquiring an image related to a specific area.

  The present invention has been made in view of the above problems, and provides a technique for acquiring an image in which a fluctuation component caused by a region other than the region is reduced and appropriately diagnosing the lung. The purpose is to do.

  In order to solve the above-described problem, the invention according to claim 1 captures a state in which the shape of the lung field of an animal including a human being changes, and temporal variation caused by the shape change of the lung field based on the respiratory cycle. A moving image including a component and a high-frequency component having a fluctuation period shorter than that of the respiratory cycle, one reference image included in the moving image, and a plurality of remaining images obtained by removing the reference image from the moving image With respect to a certain target image, detection means for detecting a corresponding attention area, which is an area corresponding to each attention area of the reference image, from each of the target images, and temporal values of pixel values of the corresponding attention area corresponding to each of the attention areas An image generation apparatus comprising: a generation unit configured to generate a corrected moving image by performing a process of reducing the high-frequency component from a change.

  The invention according to claim 2 is the image generation apparatus according to claim 1, wherein the detection means sets the region of interest from a region where a blood vessel is captured in the reference image. To do.

  The invention according to claim 3 is the image generation apparatus according to claim 1, wherein the detection unit is configured to detect the lung in each target image with respect to a first image region in which the lung field is captured in the reference image. After correcting each of the second image regions so that the second image region capturing the field matches, the region of each second image region that has the same positional relationship as each of the attention regions in the first image region Are detected as the corresponding attention areas.

  According to a fourth aspect of the present invention, there is provided a program that causes a computer to function as the image generating apparatus according to any one of the first to third aspects when executed by the computer.

  Further, the invention of claim 5 captures a state in which the shape of the lung field of an animal including a human being changes, a temporal variation component resulting from the shape change of the lung field based on the respiratory cycle, and a period of variation from the respiratory cycle. Each of the reference images is acquired for one reference image included in the moving image and a target image that is a plurality of remaining images obtained by removing the reference image from the moving image. A detection step of detecting a corresponding attention area, which is an area corresponding to the attention area, from each of the target images, and a process of reducing the high-frequency component from a temporal change in the pixel value of the corresponding attention area corresponding to each of the attention areas A generation step of generating a corrected moving image by performing the above.

  According to the present invention, for a lung whose shape changes according to the respiratory cycle, a moving image in which a high-frequency component having a shorter fluctuation cycle than the respiratory cycle can be acquired from the pixel value, so that the lung is appropriately diagnosed. Is possible.

  According to the second aspect of the present invention, since the region of interest is set from the region in the lung field where the blood flow particularly affected by blood flow is captured, the fluctuation component due to the blood flow is accurately detected from the moving image at high speed. Can be deleted well.

  According to the invention described in claim 3, since the shape of the lung field of the reference image and the target image are matched, the corresponding attention area of the target image can be easily detected.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<Overall configuration of image generating apparatus>
FIG. 1 shows a configuration common to each embodiment in which the present invention is applied to an image generating apparatus 1 capable of generating a diagnostic image of a lung of a patient in cooperation with a medical image capturing apparatus and an image generating apparatus 50 described later. It is a block diagram.

  As shown in FIG. 1, the image generation apparatus 1 has a general computer configuration in which a control unit 2, a display unit 3, an operation unit 4, an input unit 5, and a storage unit 6 are connected to a bus line 10. .

  The control unit 2 is constituted by, for example, a CPU, determines the operation of the entire image generation device 1 by executing a control program stored in the storage unit 6, gives a command to the entire image generation device 1, and will be described later. A display instruction is given to the display unit 3. The control unit 2 creates a corrected moving image as means for realizing each function described later.

  The display unit 3 is configured by a liquid crystal display, for example, and visually outputs moving image data generated by the control unit 2.

  The operation unit 4 includes a keyboard, a touch panel, a mouse, and the like, and transmits various command signals to the control unit 2 in accordance with various user operations.

  The input unit 5 inputs image data. The input unit 5 may receive image data on-line by connecting a medical image photographing device, and further input data by reading data from a portable storage medium such as a DVD or reading by a scanner. Is possible. Alternatively, an image obtained by photographing a person who is a subject of photographing (photographing subject) is stored in a file server or the like connected via a network, and the image of the subject subject of photographing is selected from a plurality of stored image data. Data may be retrieved and read. The input image data is stored in the storage unit 6.

  The storage unit 6 is configured by a storage device such as a semiconductor memory or a hard disk, and stores a program executed by the control unit 2, information necessary for executing the program, and information such as a moving image input from the input unit 5. To do.

  The medical image photographing apparatus is constituted by, for example, an X-ray photographing apparatus or the like, and photographs a predetermined part included in a built-in person of the photographing subject. The X-ray imaging apparatus performs imaging by exposing a subject to be imaged from an X-ray generation source. The exposed X-rays pass through the chest of the subject, the intensity distribution is detected, the detected X-rays are converted into analog electrical signals, and the analog electrical signals are converted into digital signals by A / D conversion. Then, it is stored on the storage device of the X-ray imaging apparatus as a moving image consisting of a plurality of still images in time series. The image stored in the storage device is transferred to the input unit 5 as necessary.

  In the following first and second embodiments, an X-ray image is used as a radiation image.

<First Embodiment>
In the first embodiment, an image that creates an image in which the influence of blood flow is reduced based on information of an image region that captures a blood vessel in the lung field region in a moving image that captures a change in the shape of the lung field based on the respiratory cycle. The generation apparatus will be described as an example.

  FIG. 2 is a diagram illustrating a functional configuration realized by the image generation apparatus 1. The image generation apparatus 1 includes a moving image acquisition unit 11, a target pixel calculation unit 12 corresponding to a detection unit, and a fluctuation component reduction unit 13 corresponding to a generation unit.

  The moving image acquisition unit 11 acquires a moving image that is input from the input unit 5 and stores the lung field region stored in the storage unit 6, and outputs the acquired moving image to the target pixel calculation unit 12. The moving image acquired here is an X-ray image obtained by imaging a lung field region of at least one respiratory cycle. One respiratory cycle (cycle) is a cycle of respiratory motion including one exhalation mode and an inspiration mode, and the inspiration mode is a mode for inhaling a breath. As the breath expires, the area of the lung field in the thorax increases and the diaphragm is pushed down. The exhalation mode is a mode for exhaling, and as the exhalation is performed, the lung field region becomes smaller and the diaphragm rises.

  The pixel-of-interest calculation unit 12 uses one moving image as a reference image, creates an image that captures blood vessels that are considered to be affected by blood flow from the reference image, The target pixel that is the target pixel is sequentially set. The set target pixel is output to the fluctuation component reducing unit 13.

  The fluctuation component reducing unit 13 detects the temporal change in the pixel value of the corresponding target pixel that is the pixel corresponding to each target pixel of the reference image in the target image that is the remaining plurality of images obtained by removing the reference image from the moving image. A corrected moving image is generated by performing processing for reducing high-frequency components.

  FIG. 3 is a flowchart illustrating a procedure in which the image generation apparatus 1 generates a corrected moving image. The image generation device 1 sequentially sets a target pixel from a region where a blood vessel is captured in a reference image, and reduces high-frequency components from temporal changes in the pixel value of the corresponding target pixel corresponding to each target pixel. By performing the processing, a corrected moving image is generated.

  In step S1, the moving image acquisition unit 11 executes the pixel-of-interest calculation unit 12 from step S2 to step S6, and the fluctuation component reduction unit 13 from step S7 to step S9. First, when the image generating apparatus 1 receives an instruction for generating a corrected moving image and transmits the instruction to the control unit 5, the process proceeds to step S1.

  In step S1, the moving image acquisition unit 11 acquires an X-ray moving image obtained by imaging a lung field region during breathing, outputs the acquired X-ray moving image to the target pixel calculation unit 12, and proceeds to step S2.

  In step S2, the pixel-of-interest calculation unit 12 selects one reference image from the input moving images based on a predetermined rule, and proceeds to step S3. Examples of the predetermined rule include a rule for selecting a predetermined frame (first) frame of a moving image as a reference image.

  In step S <b> 3, the pixel-of-interest calculation unit 12 first extracts a lung field outline (lung field shape) from the moving image in order to determine a lung field region. The contour extraction of the lung field can be performed using a method disclosed in, for example, Japanese Patent Laid-Open No. 63-240832 or Japanese Patent Laid-Open No. 2-250180.

  FIG. 4 is a diagram illustrating a part of a moving image obtained by performing contour extraction of a lung field. The horizontal axis of the figure shows the elapsed time in the respiratory cycle. In FIG. 4, the sand hatched area is a lung field area. After the contours of the lung fields are extracted from all the images constituting the moving image, the process proceeds to step S4.

  In step S4 in FIG. 3, the reference image selected in step S2 is subjected to a smoothing filter sequentially for each small area, thereby blurring the image and creating a blurred reference image.

  FIG. 5 is a diagram illustrating an example of the blurring reference image. After creating the blurring reference image, the process proceeds to step S5.

  In step S5 of FIG. 3, a blood vessel image is created from the reference image and the blurred reference image. The X-ray image is black in the vicinity of the position where the blood vessel does not exist in the lung field region, so that the X-ray image has a high pixel value. Conversely, X-ray images are white at positions where blood vessels are present, so that the X-ray image is white, that is, the pixel value of the X-ray image is low. When smoothing is performed when a blood vessel exists in a peripheral region of a certain pixel in a certain pixel in the lung field region, this pixel becomes lower than the original pixel value, and the pixel in which the blood vessel exists is It becomes higher than the original pixel value. On the other hand, when there is no blood vessel in the peripheral area of a certain pixel, even if smoothing is performed, the pixel value of this pixel is kept high like the original pixel value. Therefore, when the absolute value of the difference between the smoothed image and the original image before smoothing is taken in the lung field region, the absolute value of the difference tends to increase in the region where the blood vessel exists. It is in.

  FIG. 6 is a diagram illustrating a procedure for creating a blood vessel image. As shown in FIG. 6, the absolute value of the difference between the lung field regions of the reference image and the blurred reference image is obtained, and a blood vessel image is shown that indicates a region where an absolute value exceeding a certain threshold value appears. Non-patent literature (Akinobu Shimizu, Junichi Hasegawa, Junichiro Toriwaki, “Characteristics of Rotating Second-Order Difference Filters for Computer Diagnosis of Medical Images”, Institute of Electronics, Information and Communication Engineers, Vol.J78-D- 2, 29 to 39, 1995), a technique of extracting blood vessels using a Max-DD filter may be used. After creating the blood vessel image, the process proceeds to step S6.

  In step S6 of FIG. 3, a template centered on a blood vessel pixel that is a pixel in the blood vessel image is created from the reference image, the central pixel of the template is used as a target pixel, and a plurality of moving images other than the reference image are configured. Template matching is performed on the lung field region of the target image, and the corresponding target pixel corresponding to each target pixel of the reference image is detected from each target image.

  FIGS. 7A to 7C are diagrams showing how the corresponding target pixel is obtained by searching for a region (subregion) most similar to the rectangular template in the lung field region. 7A shows a blood vessel image, FIG. 7B shows a reference image, and FIG. 7C shows a target image. As shown in FIG. 7A, blood vessel pixels are displayed in the blood vessel image. As shown in FIG. 7B, a template centered on coordinates where the blood vessel pixels exist is created from the reference image. Next, as shown in FIG. 7C, template matching is performed on the lung field region of each target image, which is an image other than the reference image, in the moving image, and correspondence is taken for each pixel. At this time, in the lung field region of the reference image, regarding the coordinates where the blood vessel pixel does not exist, it is assumed that there is no influence of fluctuation due to blood flow, and no correspondence is taken for each pixel. After searching the sub-region, the center pixel of the sub-region is set as the corresponding target pixel.

  Here, as a means for evaluating how similar a certain sub-region is to the template, a method such as a correlation mutual method, an SSDA method (Sequential Similarity Detection Algorithm), or a Fourier transform phase correlation method can be employed. In addition, as the shape of the template, a rectangle, a circle, a cross, or the like can be used, but a rectangle is preferable from the viewpoint of reducing a calculation burden. After the target pixel calculation unit 12 detects the corresponding target pixel and outputs it to the fluctuation component reduction unit 13, the process proceeds to step S7.

  In step S7 of FIG. 3, the fluctuation component reducing unit 13 creates a fluctuation graph for each target pixel and each corresponding target pixel in the lung field region. The fluctuation graph is a graph showing how the pixel value changes in each target pixel and each corresponding target pixel in the lung field region of the moving image when time passes.

  FIG. 8 is a diagram illustrating an example in which a variation graph, which is a graph showing variations in pixel values, is created for each pixel of interest and each corresponding pixel of interest in the lung field region. The horizontal axis of the figure represents the elapsed time in the respiratory cycle, and the vertical axis of the graph represents the pixel value of the target pixel and the corresponding target pixel. When the variation graph is created, the process proceeds to step S8.

  In step S8 of FIG. 3, the high frequency component of each fluctuation graph is deleted. The high-frequency component corresponds to the frequency of the heartbeat, and shows a value several to several tens of times the respiration frequency. However, by removing the high-frequency component in the fluctuation graph of FIG. Obtain the pixel value of each pixel. As a method of removing the high frequency component, a method of applying a smoothing filter to the fluctuation graph can be considered. As another method, a power spectrum is obtained by performing FFT (Fast Fourier Transform) on the fluctuation graph, and a high-frequency component is removed by applying a low-pass filter to the power spectrum, and inverse FFT is performed. Do. The high frequency component of each fluctuation graph is deleted, and the process proceeds to step S9.

  In step S9, for each corresponding target pixel in the lung field region of the target image, the pixel value obtained by reducing the high-frequency component, which is the variation component, is substituted into the original position of each corresponding target pixel, and the pixel value of the variation component To create a target image with reduced variation components, and the process ends.

  FIG. 9 is a diagram showing a variation graph after high-frequency component removal and a pixel value with a reduced high-frequency component applied to the original moving image. The vertical axis of the graph represents the pixel value of the target pixel and the corresponding target pixel, and the horizontal axis represents the elapsed time in the respiratory cycle.

  In the present embodiment, for a lung whose shape changes due to a respiratory cycle, a moving image in which a fluctuation component caused by blood flow is reduced from a pixel value can be acquired, so that the lung can be appropriately diagnosed. . In addition, a target pixel that is considered to be affected by blood flow in the lung field region is set, and a lung whose shape changes is removed by removing a fluctuation component due to blood flow from the corresponding target pixel corresponding to the target pixel. With respect to the field, it is possible to acquire an image in which the fluctuation component due to the blood flow is removed at high speed with high accuracy.

Second Embodiment
In the second embodiment, based on the information of the image area that captures the change in the shape of the lung field based on the respiratory cycle, the pixel of interest that is the pixel of interest in the reference image and the pixel that corresponds to the pixel of interest in the target image A system that creates a moving image in which the influence of blood flow is reduced by associating with the corresponding target pixel will be described.

  FIG. 10 is a diagram illustrating a functional configuration realized by the image generation device 50 according to the second embodiment. The image generation apparatus 50 includes a moving image acquisition unit 21, a corresponding target pixel calculation unit 22 corresponding to a detection unit, and a fluctuation component reduction unit 23 corresponding to a generation unit.

  A part of the functional configuration of the second embodiment is similar to that of FIG. 2 described above, different configurations will be described, and description of similar configurations will be omitted. For example, the configuration and operation from acquiring a moving image obtained by capturing a lung field region stored in the storage unit 6 of the moving image acquiring unit 21 to outputting it to the corresponding target pixel calculating unit 22 is the same as in the first embodiment. .

  The moving image acquisition unit 21 acquires a moving image obtained by photographing the lung field region and outputs the acquired moving image to the corresponding target pixel calculation unit 22.

  The corresponding target pixel calculation unit 22 matches the lung field shape of the reference image with the lung field shape of each target image that is the remaining image obtained by removing the reference image from the moving image, and in each target image, the target pixel of the reference image Are detected as the corresponding target pixels. The detected corresponding target pixel is output to the fluctuation component reduction unit 23.

  The fluctuation component reduction unit 23 generates a corrected moving image by performing a process of reducing the high frequency component from the temporal change in the pixel value of the corresponding target pixel of each target image.

  FIG. 11 is a flowchart illustrating a procedure in which the image generation device 50 generates a corrected moving image. The image generation device 50 corrects each target image so that the lung field region in each target image matches the lung field region in the reference image, and among the target images, the positional relationship with the target pixel in the reference image is the same. Is detected as each corresponding target pixel, and a corrected moving image is generated by performing a process of reducing the high frequency component from the temporal change of the pixel value of the corresponding target pixel corresponding to each target pixel.

  In step T1, the moving image acquisition unit 21 executes the corresponding target pixel calculation unit 22 from step T2 to step T6, and the fluctuation component reduction unit 23 from step T7 to step T9. First, when the image generation apparatus 50 receives an instruction to generate a corrected moving image and transmits the instruction to the control unit 5, the process proceeds to step T1.

  In step T1, as in step S1 of FIG. 3, the moving image acquisition unit 21 acquires an X-ray moving image obtained by imaging a lung field region during breathing, and outputs the acquired X-ray moving image to the corresponding target pixel calculation unit 22 to step T2. Move.

  In step T2, as in step S2 of FIG. 3, the corresponding target pixel calculation unit 22 selects one reference image from the input moving images, and proceeds to step T3.

  In step T3, as in step S3 of FIG. 3, the corresponding target pixel calculation unit 22 extracts a lung field outline (lung field shape) from the moving image. After the lung field contour is extracted, the process proceeds to step T4.

  In step T4, the lung field region of each target image is corrected so that the lung field region of each target image matches the lung field region of the reference image.

  FIG. 12 is a diagram illustrating an image corrected so that the lung field shape of the target image matches the lung field shape of the reference image. As a technique for correcting the shift between the two images, various known techniques can be adopted, and disclosed in, for example, Japanese Examined Patent Publication No. 61-14553 and Japanese Unexamined Patent Publication No. 63-278183. Such a known method can be adopted.

  Here, FIGS. 13 to 15 are diagrams for explaining a method of matching the lung field region of the reference image with the lung field region of each target image.

  Hereinafter, a method of using a set of line segment conversion will be described in detail. Here, a predetermined point (for example, the upper left point) of each frame image is a reference point (for example, the origin), the right direction is the X axis direction, the lower direction is the Y axis direction, and the coordinate value for each pixel is It will change one by one.

  First, as schematically shown in FIGS. 13 (a) and 13 (b), a large number of lung field regions of the target image to be corrected and the lung field regions of the reference image corresponding to the lung field region of the target image. Feature points (for example, points with black circles in FIG. 10) (for example, about 200) are detected. Examples of the feature point include an inflection point, a sharp point, or a point having a predetermined gradation in the lung field region of the target image and the lung field region of the reference image.

  Next, the correspondence between an arbitrary point X in the lung field region of the target image and an arbitrary point X ′ in the lung field region of the reference image is obtained. Here, the correspondence between the point X and the point X ′ is recognized for all the pixels in the lung field region of the target image.

  Here, as shown in FIG. 14, two points P and Q included in the feature points of the lung field region of the target image detected as described above, and the feature points of the lung field region of the reference image detected as described above. Focus on two points P ′ and Q ′ corresponding to the two points P and Q.

  FIG. 15 is a diagram for explaining how to obtain the correspondence between the point X and the point X ′. In FIG. 15A, the vector from the point P ′ to the point Q ′ (P′Q ′ vector), the point C ′ to the point X ′ that coincides with the perpendicular from the line segment P′Q ′ to the point X ′. A vector (C′X ′ vector) and a vector from the point P ′ to the point C ′ (P′C ′ vector) are shown. On the other hand, in FIG. 15B, a vector from the point P to the point Q (PQ vector), a vector from the point C to the point X that coincides with a perpendicular from the line segment PQ to the point X (CX vector), and a point P A vector from point to point C (PC vector) is shown.

  Here, when the P′Q ′ vector is converted into the PQ vector on the two-dimensional plane, the ratio between the P′Q ′ vector, the C′X ′ vector, and the P′C ′ vector, the PQ vector, and the CX vector It is assumed that the point X ′ is converted to the point X so that the ratio between the vector and the PC vector is constant.

  At this time, the following expressions (1) to (3) are established between the point X and the point X ′.

  Note that u represents a scalar obtained by dividing (normalizing) the length of the line segment PC by the length of the line segment PQ, v represents the distance from the straight line PQ to the point X, and X ′ represents the coordinates of the point X ′. Is shown. Further, (X−P) represents a coordinate change value obtained by subtracting the coordinate value of the point P from the coordinate value of the point X, that is, a vector (PX vector) from the point P to the point X, and (Q−P) represents A coordinate change value obtained by subtracting the value of the coordinate of the point P from the value of the coordinate of the point Q, that is, a PQ vector, Per (QP) is perpendicular to the PQ vector and is equal in magnitude and in the same direction as the CX vector Per (Q′−P ′) is a vector perpendicular to and equal in magnitude to the P′Q ′ vector and facing the same direction as the C′X ′ vector, and P ′ is a point The coordinates of P ′ are shown.

  Therefore, for the right side of the above equation (1), the numerator indicates the inner product of the PX vector and the PQ vector, and the denominator indicates the square of the size of the PQ vector (the square of the length of the line segment PQ). Yes. Regarding the right side of the above equation (2), the numerator indicates the inner product of the PX vector and Per (Q−P), and the denominator indicates the size of the PQ vector (the length of the line segment PQ). Further, the denominator of the third term on the right side of the above equation (3) indicates the size of the P′Q ′ vector (the length of the line segment PQ).

  Here, for the lung field region of the target image, arbitrary two points among the many feature points are set as points P and Q, and for the lung field region of the reference image, two points corresponding to these two points P and Q are obtained. Points P ′ and Q ′ are set. A large number of relationships between the point X and the point X ′ are obtained with respect to an arbitrary point X in the lung field region of the target image. However, the relationship between the point X and the point X ′ obtained for the setting of the points P and Q (and the points P ′ and Q ′) does not completely match. Therefore, a weighted average is used to obtain the relationship between the point X ′ and the point X as shown by the following equations (4) to (6) for each point X. In addition, as the weighted average, for example, weighting considering the size of the PQ vector, the detection accuracy of the points P and Q, the positional relationship between the PQ vector and the point X, and the like may be used.

Here, n points P and Q are set, and in the above equations (4) to (6), among the combinations of n points P and Q, the i-th point P and Q are respectively points P. _i and Q _i are indicated. X ′ _i is a coordinate obtained by the above equation (3) with respect to a vector from the point P _i to Q _i (P _i Q _i vector). Furthermore, a, b, and p are constants, and the way of deforming the lung field region of the target image is adjusted by appropriately changing the values of a, b, and p. As for the above equation (4), X ′ represents the coordinate of the point X ′, X represents the coordinate of the point X, ω _i represents the weight defined by the above equations (5) and (6), (X ′ _i −X) represents a vector (XX ′ _i vector) from the point X to the point X _i ′.

As shown in the above equation (5), the weight ω _i is obtained by dividing the b-th power of the vector (P _i Q _i vector) from the point P _i to the point Q _i by (a + DIST _i ) and raising it to the p-th power. . It should be noted that the condition represented by the above equation (6), u <0, is a condition where the point X is positioned on a perpendicular to the extension line on the P side of the line segment PQ, and 0 ≦ u ≦ 1, the point X is a line U <1 is a condition where the point X is located above the perpendicular to the extension line on the Q side of the line segment PQ.

As shown in the above equation (4), the coordinate of the point X ′ corresponding to the point X is obtained by adding the weighted average of the XX _i ′ vector of the second term on the right side to the coordinate of the point X.

  When the relationship between the point X and the point X ′ obtained by the above equations (1) to (6) is obtained for all the pixels in the lung field region of the target image, the lung field region of the target image is the lung of the reference image. It is corrected to match the field area. Such correction of the shift between the two images is an image process for performing non-linear enlargement and reduction of the image region, and can be referred to as a process for performing non-linear scaling, for example. After correcting the lung field region of each target image, the process proceeds to step T5.

  Next, in step T5 of FIG. 11, after each pixel in the target image area having the same positional relationship as the pixel of interest in the reference image area is detected as each corresponding pixel of interest and output to the fluctuation component reduction unit 23. Then, the process proceeds to step T6.

  In step T6, as in step S7 of FIG. 3, the fluctuation component reducing unit 23 creates a fluctuation graph for each target pixel and each corresponding target pixel in the lung field region of the target image. FIG. 16 is a diagram illustrating an example in which a variation graph is created for the target pixel and each corresponding target pixel in the lung field region of the target image. The horizontal axis of the figure shows the elapsed time in the respiratory cycle.

  In step T7 of FIG. 11, the high frequency component of each fluctuation graph is deleted. Also in the present embodiment, similarly to the first embodiment, the high frequency component in the fluctuation graph of FIG. 16 is removed to create a fluctuation component reduced image.

  In step T8, for each corresponding target pixel in the lung field region of the target image, the pixel value obtained by reducing the high-frequency component that is the fluctuation component is substituted into the position of each corresponding target pixel, and the pixel value of the fluctuation component To create a target image with reduced fluctuation components.

  FIG. 17 is a diagram showing a variation graph after removal of high-frequency components and a pixel value with reduced high-frequency components applied to the original moving image. The horizontal axis of the figure shows the elapsed time in the respiratory cycle.

  In step T9 in FIG. 11, the lung shape of the deformed target image is returned to the original lung field shape, and the process ends as a corrected moving image.

  FIG. 18 is a diagram illustrating a state in which the corrected moving image is generated by returning to the original lung field shape. As a method for returning to the original shape of the lung field, the deformation parameters (a, b, p) are stored when the nonlinear scaling process as described above is performed, and the deformed target image is based on the parameters. There is a method of returning the lung field shape of the original to the lung field shape of the original target image. Further, a method of applying a non-linear scaling process again to match the lung field shape of the deformed target image with the lung field shape of the original target image may be used.

  As described above, also in the present embodiment, it is possible to acquire a moving image in which a fluctuation component due to blood flow is reduced from a pixel value for a lung whose shape changes according to a respiratory cycle. In addition, because the shape of the lung field of the reference image and the target image are matched, it is possible to easily detect the corresponding pixel of interest in the target image, and to acquire an image in which the fluctuation component due to blood flow is accurately removed It becomes.

<Modification>
As mentioned above, although embodiment of this invention was described, this invention is not limited to the thing of the content demonstrated above.

  For example, in the above-described embodiment, the case where the image generation apparatus has a configuration independent of the X-ray imaging apparatus has been described. However, each unit of the image generation apparatus may be realized by a computer provided in the X-ray imaging apparatus itself.

  In the above embodiment, the moving image to be subjected to image processing is a transmission image related to the internal structure of the human body using X-rays, but is not limited thereto. For example, various moving images that capture the internal structure of the lung using MRI (Magnetic Resonance Imaging) or the like may be used.

  In the above embodiment, the attention area of the reference image and the corresponding attention area of the target image are processed in units of pixels. However, the processing may be performed in units of a group of a plurality of pixels. .

  Note that the configurations described in the above embodiments and modifications may be replaced as appropriate as long as no contradiction arises.

  These modifications can also achieve the same effects as in the above embodiment.

2 is a block diagram illustrating a configuration common to the embodiments applied to the image generation apparatuses 1 and 50. FIG. It is a figure which shows the function structure implement | achieved by the image generation apparatus 1 which is 1st Embodiment. It is a flowchart which shows the procedure in which the image generation apparatus 1 produces | generates a correction | amendment moving image. It is a figure which shows a part of moving image which performed the outline extraction of the lung field. It is a figure which shows the example of a blurring reference image. It is a figure which shows the procedure which produces a blood-vessel image. It is a figure which shows how a corresponding attention pixel is calculated | required. It is a figure which shows the example which produced the fluctuation graph about the attention pixel in a lung field area | region, and each corresponding attention pixel. It is a figure which shows how the fluctuation value after high frequency component removal and the pixel value which reduced the high frequency component are applied to the original moving image. It is a figure which shows the function structure implement | achieved by the image generation apparatus 50 which is 2nd Embodiment. It is a flowchart which shows the procedure in which the image generation apparatus 50 produces | generates a correction | amendment moving image. It is a figure which shows the image correct | amended so that the lung field shape of a target image may correspond with the lung field shape of a reference | standard image. It is a figure for demonstrating the method to match the lung field area | region of a reference | standard image, and the lung field area | region of each object image. It is a figure for demonstrating the method to match the lung field area | region of a reference | standard image, and the lung field area | region of each object image. It is a figure for demonstrating how to obtain | require the correspondence of point X and point X '. It is a figure which shows the example which produced the fluctuation graph about the attention pixel and each corresponding attention pixel in the lung field area | region of a target image. It is a figure which shows how the fluctuation value after high frequency component removal and the pixel value which reduced the high frequency component are applied to the original moving image. It is a figure which shows a mode that it returns to the original lung field shape and produces | generates a correction | amendment moving image.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,50 Image generation apparatus 11 Moving image acquisition part 12 Attention pixel calculation part 13 Fluctuation component reduction part 21 Moving image acquisition part 22 Corresponding attention pixel calculation part 23 Fluctuation component reduction part

Claims (5)

A moving image that captures changes in the shape of the lung field of animals, including humans, and includes temporal fluctuation components due to changes in the lung field shape based on the respiratory cycle and high-frequency components that have a shorter fluctuation cycle than the respiratory cycle And a reference image included in the moving image, and a target image that is a plurality of remaining images obtained by removing the reference image from the moving image, are regions corresponding to respective attention regions of the reference image. Detecting means for detecting a corresponding attention area from each of the target images;
Generating means for generating a corrected moving image by performing a process of reducing the high-frequency component from a temporal change in a pixel value of the corresponding attention area corresponding to each of the attention areas;
An image generation apparatus comprising: The image generation apparatus according to claim 1,
The image generation apparatus according to claim 1, wherein the detection unit sets the region of interest from a region where a blood vessel is captured in the reference image. The image generation apparatus according to claim 1,
The detection means corrects each second image region so that the second image region capturing the lung field in each target image matches the first image region capturing the lung field in the reference image. After that, an image generation apparatus characterized by detecting, as the corresponding attention area, an area having the same positional relationship as the attention area in the first image area among the second image areas. By being executed by a computer
A program that causes the computer to function as the image generation apparatus according to any one of claims 1 to 3. A moving image that captures changes in the shape of the lung field of animals, including humans, and includes temporal fluctuation components due to changes in the lung field shape based on the respiratory cycle and high-frequency components that have a shorter fluctuation cycle than the respiratory cycle And a reference image included in the moving image, and a target image that is a plurality of remaining images obtained by removing the reference image from the moving image, are regions corresponding to respective attention regions of the reference image. A detection step of detecting a corresponding attention area from each of the target images;
Generating a corrected moving image by performing a process of reducing the high-frequency component from a temporal change in a pixel value of the corresponding attention area corresponding to each of the attention areas;
An image generation method comprising:

Images