JP6551252B2 - Image processing apparatus, program and radiation imaging apparatus - Google Patents

Description

  The present invention relates to an image processing apparatus, a program, and a radiation imaging apparatus that improve the visibility of a part of a radiation image.

  FIG. 25 shows a radiation image taken by a radiation imaging apparatus. When it is desired to observe the lung field of a subject that is reflected in such a radiation image, it is necessary to adjust the contrast of the lung field to perform image processing to improve the visibility of the lung field.

  In addition to the lung field, various portions of the subject such as the bone portion of the subject are reflected in the radiographic image. The bones of the subject appear dark in the radiation image because they are difficult to transmit radiation. Furthermore, the outside of the contour of the subject that is reflected in the radiation image is a portion in which air is reflected in the portion where the object is not reflected. The part where the subject is not reflected is bright in the radiographic image because there is nothing to transmit the radiation. In the radiation image, the lung field is brighter than the bone part of the subject and darker than outside the contour of the subject which is a part where air is reflected.

  The lung field of the radiographic image appears to be filled with gray with a poor overall density. This is because pixels located in the lung field reflected in the radiation image have similar pixel values.

  If the contrast adjustment is performed on the entire radiation image in order to enhance the visibility of the lung field, the contrast adjustment is performed including the bone part included in the radiation image and the part where air is reflected. Such contrast adjustment may increase the visibility of the radiation image as a whole, but the improvement in visibility can not be seen as much as seen in the lung field. The lung field after contrast adjustment remains poorly light as a whole. In order to express the density of the lung field, low pixel values are used to represent the bones of the subject, and high pixel values are used to represent the portion where air is reflected, and the lungs This is because the field has to be expressed by the remaining moderate pixel values.

  Therefore, conventionally, a method of performing contrast adjustment only on the lung field in a radiation image has been devised. According to this method, the lung field can be expressed in a variety of colors, so the visibility of the lung field is reliably increased. In this method, trimming for extracting a lung field in a radiographic image is performed, and contrast adjustment is performed for a trimmed image in which the lung field is greatly captured. In the trimmed image, the bone portion of the dark subject and the portion where the bright air is reflected are excluded, so that the trimmed image is not affected by these portions.

  FIG. 26 shows a lung field search process for searching for a lung field to be an object of contrast adjustment from a radiation image. In this lung field search process, based on the distribution of pixel values of the object image, a portion where the lung field is reflected is searched out from the radiation image, and a region surrounding this portion is designated. From the viewpoint of performing contrast adjustment only on the lung field, it is desirable that the region be as narrow as possible while including the entire lung field. An area found by the lung field search process is a trimming image to be subjected to the contrast adjustment (see, for example, Patent Document 1 as an example of the image process).

JP, 2007-300966, A

  However, in the contrast adjustment of the conventional configuration, no consideration is given to the direction in which the object appears.

  The lung field search processing of the conventional configuration is premised on that the upper part of the subject is photographed in the direction positioned on the upper side of the image. In such an image, the right lung of the subject is captured on the right side of the image, and the left lung of the subject is captured on the left side. And the area under the right lung and the left lung in the image is not the lung field. The conventional lung field search processing is configured to search for a lung field subjected to such reflection from among images.

  FIG. 27 shows a state in which a conventional lung field search process is performed on a radiographic image captured with the subject lying sideways. In this radiographic image, the subject image appears sideways, so the right lung image that should appear on the right side of the image appears at the top of the image, and the left lung image that should appear on the left side of the image appears at the bottom of the image . Such a radiation image is often taken when the subject of a child is taken, but this is not necessarily the case even with normal taking.

  When a lung field search process is performed on a radiation image as shown in FIG. 27, a phenomenon may occur in which the right lung image appearing at the upper part of the image is misidentified as the lung field. Below the right lung image is the left lung image, which must be included in the lung field, but as a lung field search process, the bright part of the right lung image at the top of the image is displayed. It can only be recognized that there is no lung field below it. Therefore, according to the conventional configuration, only the right lung image is subjected to contrast adjustment. As described above, in the conventional image processing, there is a problem that the contrast adjustment can not be properly performed depending on the radiation image.

  The present invention has been made in view of such circumstances, and the object of the present invention is to provide a lung field that is surely reflected in an image even if the subject image is rotated and reflected in the radiation image. It is an object of the present invention to provide an image processing apparatus capable of reliably performing contrast adjustment with excellent visibility on a lung field.

The present invention has the following configuration in order to solve the above-described problems.
In other words, the image processing apparatus according to the present invention is configured to adjust the luminance of a portion corresponding to a lung field in a radiographic image in which a subject image is captured, and to convert pixel values belonging to each pixel column of the radiographic image into Profile generation means for generating a profile by summing or averaging each time, center position calculation means for calculating the center position of the center of the subject image reflected in the radiation image, and the profile value is minimized If the minimum position is near the center position of the subject image , it is determined that the subject image is reflected vertically in the radiation image, and the minimum position of the profile is not near the center position. lung search processing for searching for the rollover determination means determines that the subject image is overturned in the radiographic image, the lung field visible on captured radiographic image based on the result of the determination in the case It is characterized in that it comprises means.

  [Operation / Effect] According to the present invention, even if the subject image is rotated and reflected in the radiation image, the lung field that is surely reflected in the image is found, and contrast adjustment with excellent visibility is performed. An image processing apparatus that can be reliably executed on a field can be provided. That is, when the minimum position of the profile generated by summing or averaging the pixel values belonging to each pixel row of the radiation image for each pixel row is apart from the center position of the object image, the object image is within the radiation image And a rollover judging means for judging that the rollover has occurred. Since it is possible to search for the lung field while knowing how the lung field is reflected in the radiation image by the rollover judging means, it is possible to reliably perform contrast adjustment with excellent visibility on the lung field.

  In the image processing apparatus described above, the image processing apparatus further includes an image rotation unit that performs rotation processing for rotating the radiation image on the radiation image determined to be overturned by the rollover determination unit, and the lung field search processing unit rotates It is more desirable to perform a lung field search on the processed radiographic image.

  [Operation and Effect] The above-described configuration is for describing a more specific configuration of the present invention. The radiation image determined to be overturned by the rollover judging means is rotated by the image rotating means and used for searching for a lung field. Therefore, according to the present invention, even if the subject image rotates and is reflected in the radiation image, the lung field search can be performed with the subject image corrected in the correct direction.

  In the above-described image processing apparatus, the lung field search processing means searches the lung field for the radiation image determined by the rollover judging means to be that the subject image is reflected vertically in the radiation image. More desirable.

  [Operation and Effect] The above-described configuration is for describing a more specific configuration of the present invention. That is, according to the apparatus of the present invention, when the object image is vertically reflected in the radiation image, the lung field search processing is executed without rotating the image.

  Further, in the above-described image processing apparatus, an outline extraction means for generating an outline extraction image by extracting an outline of an image reflected in a radiation image determined to be lying down by the rollover judgment means; A right / left determination means for comparing the pixel value variation in the right region with the pixel value variation in the left region and determining that the region with the larger variation hits the upper part of the subject image is provided. It is more desirable that the means rotate the radiation image so that the upper part of the object image faces the upper side of the radiation image according to the determination result of the left / right determination means.

  [Operation and Effect] The above-described configuration is for describing a more specific configuration of the present invention. By providing left / right determination means for determining in which direction the subject image is overturned with respect to the radiation image determined to be overturned by the overturn determination means, the upper part of the subject image can be reliably determined. The radiation image can be rotated so that it faces the top of the radiation image.

  Further, in the above-described image processing apparatus, the outline extraction image is extracted by extracting the outline of the image to be reflected in the radiation image determined to be the one in which the subject image is vertically reflected in the radiation image by the rollover judging means. The contour extraction means to be generated, the pixel value variation in the upper region of the contour extraction image, and the pixel value variation in the lower region are compared, and the region with the larger variation hits the upper part of the subject image It is more preferable that the radiation image be rotated so that the upper and lower determination means determine the object and the upper part of the object image on the lower side of the radiation image is directed to the upper side of the radiation image.

  [Operation and Effect] The above-described configuration is for describing a more specific configuration of the present invention. As long as the apparatus is provided with upper and lower determination means for determining whether or not the subject is turned upside down with respect to a radiation image determined to be a subject image vertically reflected by the rollover determination means The radiation image can be rotated so that the upper part of the object image faces the upper side of the radiation image.

  According to the present invention, even if the subject image is reflected in the radiation image by rotating, the lung field reliably reflected in the image is searched for, and the contrast adjustment with excellent visibility is assured for the lung field. Can provide an image processing apparatus that can be executed. That is, when the minimum position of the profile generated by summing or averaging the pixel values belonging to each pixel row of the radiation image for each pixel row is apart from the center position of the object image, the object image is within the radiation image And a rollover judging means for judging that the rollover has occurred. The radiation image determined to be overturned by the overturn determination means is rotated by the image rotation means and used for searching for a lung field.

FIG. 7 is a schematic view that roughly explains the operation of the image processing apparatus according to the first embodiment; FIG. 6 is a schematic view for explaining an operational feature of the image processing apparatus according to the first embodiment. FIG. 2 is a functional block diagram for explaining the configuration of the image processing apparatus according to the first embodiment. FIG. 6 is a schematic diagram illustrating profile generation processing according to the first embodiment. FIG. 6 is a schematic diagram illustrating profile generation processing according to the first embodiment. FIG. 6 is a schematic diagram illustrating profile generation processing according to the first embodiment. FIG. 7 is a schematic view illustrating a profile generation process according to the first embodiment. FIG. 6 is a schematic diagram illustrating characteristics of a profile according to the first embodiment. FIG. 6 is a schematic view for explaining problems of the profile according to the first embodiment. FIG. 7 is a schematic view illustrating center position calculation processing according to the first embodiment; FIG. 7 is a schematic view illustrating center position calculation processing according to the first embodiment; FIG. 7 is a schematic view illustrating center position calculation processing according to the first embodiment; FIG. 6 is a schematic view illustrating the relationship between the center position and the profile according to the first embodiment. FIG. 6 is a schematic view illustrating rollover determination processing according to the first embodiment; FIG. 6 is a schematic view illustrating rollover determination processing according to the first embodiment; FIG. 5 is a schematic diagram illustrating image rotation processing according to the first embodiment. FIG. 6 is a schematic view for explaining problems of the profile according to the first embodiment. It is a schematic diagram explaining the left-right determination process which concerns on Example 1. FIG. It is a schematic diagram explaining the left-right determination process which concerns on Example 1. FIG. It is a schematic diagram explaining the left-right determination process which concerns on Example 1. FIG. FIG. 6 is a schematic view for explaining problems of the profile according to the first embodiment. It is a schematic diagram explaining the left-right determination process which concerns on Example 1. FIG. It is a schematic diagram explaining the left-right determination process which concerns on Example 1. FIG. It is a schematic diagram explaining the left-right determination process which concerns on Example 1. FIG. It is a schematic diagram explaining the image processing apparatus of conventional structure. It is a schematic diagram explaining the image processing apparatus of a conventional structure. It is a schematic diagram explaining the problem of the image processing apparatus of conventional structure.

  Subsequently, an embodiment according to the present invention will be described. When the image processing apparatus 1 according to the present invention inputs a chest X-ray image (original image P0) of a subject imaged by an X-ray imaging apparatus as shown in FIG. 1, the lung field is reflected anywhere in the original image P0. It is configured to recognize the presence or absence of a lung field, to set a lung field, and to output an image subjected to color tone correction to the set lung field. The image processing apparatus according to the present invention is configured to perform luminance adjustment on a portion corresponding to the lung field in the original image P0 in which the subject image is reflected.

  According to the present invention, in addition to such conventional functions, as shown in FIG. 2, when an original image P0 in which an object image is reflected and reflected is inputted, the object image is recognized to be inverted and the object is The original image P0 is rotated so that the orientation of the image matches the orientation of the original image P0. The rotation of the original image P0 is performed prior to lung area recognition and tone correction. Therefore, when the original image P0 in which the subject is rolled up is input to the apparatus according to the present invention, the original image P0 is rotated by 90 ° to correct the rollover of the subject and then the lung field region. Settings and tone corrections will be made. The present invention is most characterized by the function of rotating the original image P0.

  FIG. 3 is a functional block diagram showing the entire image processing performed by the image processing apparatus 1. Referring to FIG. 3, the profile generation unit 11, the center position calculation unit 12, the rollover determination unit 13, the left / right determination unit 14, and the image rotation unit 16 perform an operation of rotating the original image P0 described in FIG. This operation is not performed on the original image P0 in which the subject is not reflected in the vertical direction as described in FIG.

  The upper / lower determination unit 15 is configured to rotate the subject image by 180 ° with respect to an original image P0 in which the object is reflected upside down among the original images P0 in which the object is reflected in the vertical direction. . The lung field search processing unit 17 is configured to search for where the lung field in the subject image is on the original image P0 and to determine the lung field region. The brightness adjustment unit 18 is configured to perform contrast adjustment (color tone correction) on the lung region.

  The profile generator 11 corresponds to the profile generator of the present invention, and the center position calculator 12 corresponds to the center position calculator of the present invention. The rollover judging unit 13 corresponds to the rollover judging unit of the present invention, and the image rotating unit 16 corresponds to the image rotating unit of the present invention. The lung field search processing unit 17 corresponds to the lung field search processing means of the present invention.

  The apparatus according to the present invention determines whether the subject image is reflected vertically or horizontally in the original image P0, and determines whether to rotate the original image P0 according to the determination result. The configuration is characteristic. How to determine the presence or absence of rollover of the object image will be described in order, as the profile generation unit 11, the center position calculation unit 12, and the rollover determination unit 13 are related.

<Operation of Profile Generation Unit 11>
FIG. 4 shows the operation of the profile generation unit 11. The profile generation unit 11 averages the pixel values of each pixel belonging to the target pixel column in which the pixels are arranged in the vertical direction having a width of one pixel in the original image P0 configured by arranging the pixels vertically and horizontally. Calculate the average value. In the description of FIG. 4, the average value Ave (X1) is calculated for the pixel column related to the left end position X1 in the original image P0. The profile generation unit 11 repeatedly calculates the average value Ave while moving the target pixel column to the right by one pixel in the horizontal direction, and calculates up to the pixel column related to the right end position in the original image P0. Then, as illustrated in FIG. 5, the profile generation unit 11 generates a profile in which the average value Ave and the position of the target pixel column related to the average value Ave are associated. Therefore, this profile is a laterally elongated profile for the original image P0. As described above, the profile generation unit 11 generates a profile by summing or averaging the pixel values belonging to each pixel column of the original image P0 for each pixel column.

  In the above description, the profile generation unit 11 sequentially calculates the average value Ave from the left end to the right end of the original image P0, but the order of calculation does not have to be this procedure. This situation also applies to FIGS. 6 and 7 described later.

  In the explanation of FIG. 4 and FIG. 5, the object image is reflected vertically in the original image P0. By the way, the original image P0 in which the object image as described in FIG. 2 is turned over may be input to the apparatus of the present invention. FIG. 6 shows how the profile generation unit 11 generates a profile for the original image P0 in which the object image is imprinted sideways. Since the original image P0 is defined as which is the top, the profile generation unit 11 averages the pixel columns related to the left end position X1 in the original image P0 while the subject image is lying sideways. The value Ave (X1) is calculated, and the calculation of the average value Ave is repeated while moving the pixel-of-interest row by one pixel in the horizontal direction. Then, a profile as shown in FIG. 7 is generated.

  FIG. 8 illustrates how the profile changes depending on whether the subject image is rolled over. The left side of FIG. 8 shows the case where a vertically oriented object is reflected in the original image P0. It can be seen that the profile in this case is concave at the center. This dent is located on the original image P0 between the bright region derived from the bright left lung and the bright region derived from the bright right lung, and the value of this portion is the pixel of the pixel located in the dark spine It is the average value calculated by averaging the values.

  On the other hand, the right side of FIG. 8 shows a case in which the subject rolls over in the original image P0. In the profile in this case, no depression appears in the central portion, but only one broad bright region appears in the central portion. The influence of pixel values of pixels located in the dark spine on the original image P0 on the profile is dispersed over the entire profile, and is not concentrated on a specific part of the profile. Therefore, the lowest value in the profile at this time is at the end where the bright lung field is not located on the original image P0.

  Therefore, when the lowest value of the profile appears at the center of the profile, it can be said that the vertically oriented subject is reflected in the original image P0. When the lowest value of the profile appears at the end of the profile, it can be said that the subject that has rolled over is reflected in the original image P0. In practice, the present invention uses this principle to determine the presence or absence of the rollover of the object incorporated in the original image P0.

  However, misjudgment may occur when it is determined whether or not the subject image rolls over only at the position of the lowest value of the profile. FIG. 9 explains the circumstances. When the object image is reflected to an end on the original image P0, the center of the object image and the center of the original image P0 are separated. If the profile is generated over the entire area of the original image P0, the lowest value of the profile will appear at the end of the profile. As a result, the original image P0, in which the object image that is not originally rolled over, is misidentified as being overturned, and the image is rotated. Such a phenomenon is particularly likely to occur on the original image P0 obtained by photographing an infant.

  The present invention is designed to calculate the center of the object image reflected in the original image P0 for the purpose of preventing such false recognition. The center position calculation unit 12 executes this operation.

<Operation of Center Position Calculation Unit 12>
The center position calculation unit 12 is configured to calculate the center of the object image reflected in the original image P0. The center position calculation unit 12 first performs edge enhancement processing on the original image P0 to generate an edge enhanced image. The edge enhancement processing is performed by applying a differential filter such as a Sobel filter to the original image P0. If the center of the subject image is directly obtained with respect to the original image P0, the result differs depending on the pixel value of the pixel located in the external region of the subject image. If the edge enhancement processing is performed, the structure itself reflected on the original image P0 can be raised, so that the center of the subject image can be calculated reliably.

FIG. 11 shows a state in which the center position calculation unit 12 calculates the center of gravity G of the subject image with respect to the edge-enhanced image of the original image P0 that captures the subject in the vertical direction.
The coordinates (Gx, Gy) on the edge-emphasized image of the center of gravity G are determined as follows.
Gx = ΣI (x) · x / ΣI (x)
Gy = ΣI (y) · y / ΣI (y)
Here, x represents the position of the edge enhanced image in the horizontal direction, and I (x) represents the sum of the pixel values of the pixel at the position x of the edge enhanced image.
Further, y represents the position in the vertical direction of the edge-enhanced image, and I (y) represents the total value of the pixel values at the position y of the edge-enhanced image. In the present invention, the coordinates of the center of gravity G are used for calculation only in the horizontal direction, so it is not always necessary to calculate Gy. However, for convenience of explanation, the description will be made assuming that the position of the center of gravity G is obtained.

  In FIG. 11, since the edge-enhanced image of the subject is reflected throughout the edge-enhanced image, the position of the center of gravity G appears near the center of the image. The position of the center of gravity G indicates the center of the object image on the original image P0.

  FIG. 12 shows a state in which the center position calculation unit 12 calculates the center of gravity G of the object image with respect to the edge-emphasized image of the original image P0 with the object being turned over. In FIG. 12, since the edge-enhanced image of the subject is reflected over the entire edge-enhanced image, the position of the center of gravity G appears near the center of the image. The position of the center of gravity G indicates the center of the object image on the original image P0. As described above, the center position calculation unit 12 calculates the center position (center of gravity G) which is the position of the center of the subject image reflected in the original image P0.

  FIG. 13 shows the relationship between the center of gravity G and the profile. The left side of FIG. 13 illustrates the positional relationship in the case where the original image P0 vertically imprints the subject, and the minimum point P at which the profile is minimum is close to the position of the center of gravity G. The right side of FIG. 13 illustrates the positional relationship in the case where the original image P <b> 0 projects the object of lying down, and the minimum point P with the smallest profile is apart from the position of the center of gravity G. That is, by examining the positional relationship between the minimum point P and the center of gravity G, it is possible to know whether or not the subject on the original image P0 is lying sideways.

<Operation of Rollover Judgment Unit 13>
The rollover judging unit 13 is configured to judge the presence or absence of rollover of the object image based on such a principle. A profile of the original image P0 is sent from the profile generation unit 11 to the rollover judging unit 13, and coordinates indicating the position of the center of gravity G are sent from the center position calculation unit 12.

  FIG. 14 describes the rollover judging process performed by the rollover judging unit 13. That is, as shown on the left side of FIG. 14, the rollover judging unit 13 sets the position of the center of gravity G in the lateral direction within the predetermined range R on the basis of the lateral position of the minimum point P of the profile as shown on the left side of FIG. It is determined that the subject image of (1) is not rolled over, and is reflected in the vertical direction.

  In addition, as shown on the right side of FIG. 14, the rollover judging unit 13 places the original point P 0 on the original image when the lateral position of the center of gravity G does not belong to the predetermined range R with reference to the lateral position of the minimum point P of the profile. It is determined that the subject image of is overturned.

  Although the rollover judging unit 13 judges the position of the center of gravity G based on the minimum point P, the present invention is not limited to this configuration. The rollover may be determined based on whether the position of the minimum point P is within a predetermined range with reference to the center of gravity G. In any case, the rollover judging unit 13 turns over the object image in the original image P0 when the minimum point P, which is the position where the value of the profile is the minimum, is away from the center position of the object image. If the minimum point P of the profile is in the vicinity of the center position, it is determined that the object image is reflected vertically in the original image P0.

  As described above, if the presence or absence of the rollover of the object image is determined by comparing the position of the minimum point P of the profile and the position of the center of gravity G, as shown in FIG. Even in the case of biased reflection, correct determination can be made. When the object image is reflected toward the end on the original image P0, the position of the center of gravity G is also deviated to the end so as to follow the object image, so the minimum point P of the profile is the center of the object image It is because it will be surely grasped whether it is in or at the end.

  If the rollover judging unit 13 judges that the original image P0 is lying over, as shown in FIG. 16, the original image P0 is rotated by 90 ° by the image rotation process, and thus the object image on the original image P0 Is corrected so that the upper part (head) is above the image and the abdomen is below the image. Such image rotation processing is executed by the image rotation unit 16. The image rotation unit 16 performs rotation processing to rotate the original image P0 with respect to the original image P0 determined to be a rollover by the rollover determination unit 13.

  However, with this operation, it is impossible to determine which direction the original image P0 should be rotated by 90 °. The rollover determination unit 13 can determine whether or not the subject image on the original image P0 is rollover. If the original image P0 is rotated 90 ° clockwise, it can be distinguished whether the upper part of the subject image is on the image or if it is rotated 90 ° counterclockwise, the upper part of the subject image is on the image. is not. FIG. 17 illustrates such a situation. FIG. 17 shows an original image P0 in which the subject image is turned to the left and to the left, and is an image in which the subject image on the right side of FIG. 8 is turned in the opposite direction. The profile generated by the profile generation unit 11 based on the original image P0 of FIG. 17 is similar to the profile of the object image on the right side of FIG. Therefore, it is difficult to determine whether the upper part of the subject image corresponds to the left or right direction of the image based on the profile.

  Therefore, the apparatus according to the present invention is configured to determine whether the object image is turned to the right or to the left. This determination is performed by the left / right determination unit 14.

<Operation of left / right determination unit 14>
The left / right determination unit 14 sends the determination result of the rollover determination unit 13. When it is determined that the original image P0 is rollover, the left / right determination unit 14 first performs edge enhancement processing on the original image P0 to generate an edge enhanced image, as shown in FIG. The edge enhancement processing is performed by applying a filter such as a Laplacian filter to the original image P0. The edge-enhanced image generated at this time is referred to as a left / right determination image in distinction from the image described in FIG.

  The left / right determination unit 14 sets two regions of a right region, which is a region corresponding to the right half of the image on the left / right determination image, and a left region, which corresponds to the left half of the image. That is, the right area and the left area are set by vertically dividing the left and right discrimination image into two equal parts. Then, the left / right determination unit 14 calculates the standard deviation σ (L) of the pixel values of the pixels belonging to the right area and the standard deviation σ (R) of the pixel values of the pixels belonging to the left area.

  FIG. 19 shows an image for left / right determination according to an original image P0 in which a subject image in which the upper part (head) is turned to the right side of the image is copied. The standard deviation σ (R) calculated from this image is larger than the standard deviation σ (L). This is because the lung field of the subject is biased to the right of the image. The subject's lungs are low density and easy to pass x-rays. The lungs are protected by ribs. Since the ribs are bones, they are dense and hard to pass X-rays. Therefore, the lung field on the original image P0 is an image in which dark and bright portions are arranged in a stripe. When edge enhancement processing is performed on such an original image P0, variation in pixel values of pixels located in the lung field of the obtained left and right determination image becomes large, and becomes small in portions other than the lung field. When the lung field of the subject is biased to the right side of the image, the standard deviation σ (R) indicating the variation in the pixel value of the pixel belonging to the right region increases, and the standard indicating the variation in the pixel value of the pixel belonging to the left region. The deviation σ (L) becomes smaller. Since most of the lung field should be biased to the upper part of the subject image, the right region of the original image P0 is to hit the upper part of the subject image. If the standard deviation σ (R) is larger than the standard deviation σ (L) based on such a principle, the left / right determination unit 14 turns over the object image on the original image P0 so that the upper part is on the right side of the image. It is determined that

  FIG. 20 shows a left / right determination image related to the original image P0 in which the subject image that has been laid down so that the upper part is on the left side of the image is captured. The standard deviation σ (L) calculated from this image is larger than the standard deviation σ (R). This is because the lung field of the subject is biased to the left of the image. Since most of the lung field should be biased to the upper part of the subject image, the left region of the original image P0 is to hit the upper part of the subject image. If the standard deviation σ (L) is larger than the standard deviation σ (R) based on such a principle, the left / right determination unit 14 turns over the object image on the original image P0 so that the upper part is on the left side of the image. It is determined that

  As described above, the left / right determining unit 14 detects the outline of the image to be captured in the original image P0 that is determined to be falling over by the rollover determining unit 13 in the right region of the left / right determination image. By comparing the variation of the pixel values with the variation of the pixel values in the left region, it is determined that the region having the larger variation corresponds to the upper part of the object image.

  The determination result of the left / right determination unit 14 is sent to the image rotation unit 16. The image rotation unit 16 rotates the original image P0 based on the determination result so that the upper side of the subject image is on the image. The image rotation unit 16 rotates the original image P0 according to the determination result of the left / right determination unit 14 so that the upper part of the subject image faces the upper side of the original image P0.

  By the way, just because the rollover judging unit 13 does not detect rollover of the subject image, the direction of the subject image is not necessarily suitable for diagnosis. FIG. 21 explains this situation. FIG. 21 shows an original image P0 in which the subject image is turned upside down, and is an image in which the subject image on the right side of FIG. 8 is turned in the opposite direction. The profile generated by the profile generation unit 11 based on the original image P0 of FIG. 21 is the same as the profile of the object image on the left side of FIG. In the rollover judging unit 13, it is difficult to judge whether the object image on the original image P0 is turned upside down.

  Therefore, the apparatus according to the present invention is configured to determine whether the upper side of the object image is upward of the image or the upper side of the object image is downward of the image. This determination is performed by the upper / lower determination unit 15.

<Operation of Upper and Lower Judgment Unit 15>
The up / down determination unit 15 sends the determination result of the rollover determination unit 13. If it is determined that the original image P0 has the object image vertically transferred, the upper / lower determination unit 15 first performs edge enhancement processing on the original image P0 to generate an edge-emphasized image, as shown in FIG. Generate The edge enhancement processing is performed by applying a filter such as a Laplacian filter to the original image P0. The edge-enhanced image generated at this time is referred to as an up / down determination image in distinction from the image described in FIG.

  The up / down determination unit 15 sets two areas, an upper area that corresponds to the upper half of the image and a lower area that corresponds to the lower half of the image on the upper / lower determination image. That is, the upper area and the lower area are set by dividing the upper and lower discrimination image horizontally into two. Then, the upper / lower determination unit 15 calculates the standard deviation σ (a) of the pixel values of the pixels belonging to the upper area and the standard deviation σ (b) of the pixel values of the pixels belonging to the lower area.

  FIG. 23 shows the upper / lower determination image according to the original image P0 in which the object image turned upside down so that the upper part is below the image is copied. The standard deviation σ (b) calculated from this image is larger than the standard deviation σ (a). This is because the lung field of the subject is biased to the lower side of the image. Since the majority of the lung field should be biased to the upper part of the subject image, the lower region of the original image P0 is the upper part of the subject image. When the standard deviation σ (b) is larger than the standard deviation σ (a) based on such a principle, the upper / lower determination unit 15 sets the object image on the original image P0 so that the upper part is on the lower side of the image. It is determined that it has been turned over.

  FIG. 24 shows the upper / lower determination image according to the original image P0 in which the object image is captured so that the upper part is on the upper side of the image. The standard deviation σ (a) calculated from this image is larger than the standard deviation σ (b). This is because the lung field of the subject is biased to the upper side of the image. Since most of the lung field should be biased to the upper part of the subject image, the upper region of the original image P0 is to be hit on the upper part of the subject image. When the standard deviation σ (a) is larger than the standard deviation σ (b) based on such a principle, the upper / lower determination unit 15 reflects the object image on the original image P0 so that the upper part is on the upper side of the image. It is determined that it is crowded.

  As described above, the up / down determination unit 15 extracts the contour of the image reflected in the original image P0 determined by the rollover determination unit 13 that the subject image is reflected vertically in the original image P0. Compare the variation of the pixel value in the upper area of the generated image for vertical determination with the variation of the pixel value in the lower area, and determine that the area with the larger variation corresponds to the upper part of the subject image. Do.

  The determination result of the upper / lower determination unit 15 is sent to the image rotation unit 16. The image rotation unit 16 rotates the original image P0 that includes the subject image turned upside down so that the upper part is on the lower side of the image by 180 °. Further, the image rotation unit 16 does not rotate the original image P0 in which the object image is captured on the original image P0 so that the upper part is on the upper side of the image. That is, the image rotation unit 16 rotates the original image P0 so that the upper part of the object image located below the original image P0 faces the upper side of the original image P0.

  The original image P 0 is sent from the image rotation unit 16 to the lung field search processing unit 17. Since this original image P0 is an image after the rotation processing, the object image is reflected so that the upper part is on the upper side of the image. The lung field search processing unit 17 searches for a lung field from the original image P0. The lung field search method at this time is based on the premise that the upper side of the subject image is reflected on the upper side of the original image P0. That is, the lung field search processing unit 17 searches for the neck of the subject in the upper part of the original image P0 and operates as if a lung field exists below it, or the left end of the left lung at the left end of the original image P0. However, it operates as if the end of the right lung exists at the right end, or operates as if the end of the left lung or the right lung is located below the original image P0. Since the subject image is reflected in the original image P0 so that the upper part is on the upper side of the image, the lung field search processing unit 17 can reliably search for the lung field without misidentification. The lung field search processing unit 17 searches for a lung field to be captured in the original image P0 after rotation processing. The lung field search processing unit 17 executes a lung field search on the original image P0 determined by the rollover judging unit 13 that the subject image is reflected vertically in the original image P0.

  Data indicating where the lung field region searched by the lung field search processing unit 17 corresponds to the original image P0 is sent to the luminance adjustment unit 18. The brightness adjustment unit 18 performs color tone correction on the lung field region to improve the visibility of the lung field. The brightness adjusting unit 18 may generate a trimmed image by performing a trimming process that extracts the lung field region after the color correction and makes an independent image. The operation of the device of the present invention is completed through such an operation.

  As described above, according to the present invention, even if the subject image is rotated and reflected in the original image P0, the lung field that is surely reflected in the image is found, and contrast adjustment with excellent visibility is performed. It is possible to provide the image processing apparatus 1 that can be reliably executed on the lung field. That is, when the minimum point P of the profile generated by summing or averaging pixel values belonging to each pixel row of original image P0 for each pixel row is apart from the center position of the object image, the object image is the original A rollover judging unit 13 is provided that determines that the image P0 is a rollover. The original image P0 determined to be rollover by the rollover judging unit 13 is rotated by the image rotating unit 16 and used for searching the lung field. Therefore, according to the present invention, even if the subject image is rotated and reflected in the original image P0, the subject image can be searched for the lung field in a state in which the subject image is corrected in the correct direction, so that the visibility is excellent. Contrast adjustment can be reliably performed on the lung field.

  The present invention is not limited to the above-described embodiment, and the following modified embodiments are possible.

  (1) The present invention can also be applied to a radiation imaging apparatus on which the above-described image processing apparatus 1 is mounted.

  (2) The image processing apparatus 1 according to the present invention is also realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.

  (3) According to the above-described embodiment, the profile generation unit 11 is configured to generate a profile by calculating an average value for each pixel row constituting the original image P0, but the present invention It is not limited to. The profile generation unit 11 may repeat the operation of summing the pixel values of the pixels belonging to the pixel column, and generate the profile by arranging the calculated total value according to the position of the pixel column.

  (4) According to the above embodiment, the left / right discrimination unit 14 and the upper / lower discrimination unit 15 calculate the variation of the pixel value by calculating the standard deviation, but the present invention is not limited to this configuration. The left / right determination unit 14 and the up / down determination unit 15 may operate by calculating an index indicating variation other than the standard deviation.

  (5) According to the above-described embodiment, the left / right determination unit 14 sets the right region and the left region by vertically dividing the left / right determination image into two equal parts, but the present invention is limited to this configuration. Absent. Each area can be set so that the right and left areas do not cover the entire left / right discrimination image. Also, the widths of the right area and the left area may be made different.

  (6) According to the above-described embodiment, the upper / lower determination unit 15 sets the upper region and the lower region by dividing the upper / lower determination image into two equal parts, but the present invention is limited to this configuration. Absent. It is also possible to set each area so that the upper and lower areas do not cover the entire upper / lower discrimination image. Further, the widths of the upper region and the lower region may be different.

11 Profile Generator (Profile Generator)
12 Center position calculation unit (center position calculation means)
13 Rollover judgment unit (rollover judgment unit)
16 Image rotating unit (image rotating means)
17 Lung field search processing unit (Lung field search processing means)

Claims (7)

An image processing apparatus for performing a brightness adjustment in a portion corresponding to lung field in ray image release the subject image is fancy-through,
Profile generation means for generating a profile by summing or averaging pixel values belonging to each pixel row of a radiation image for each pixel row;
Center position calculating means for calculating a center position which is a position of a center of an object image reflected in the radiation image;
Determines that minimum position value of the profile is a position that is the minimum is crowded-through vertically by the subject image within the radiographic image when in the vicinity of the front Symbol central position of the object image Rollover judging means for judging that the object image is rollover in the radiation image when the minimum position of the profile is not in the vicinity of the central position ;
An image processing apparatus comprising: lung field search processing means for searching for a lung field to be captured in the radiation image based on a result of the determination. In the image processing apparatus according to claim 1,
An image rotation unit configured to execute rotation processing for rotating the radiation image with respect to the radiation image determined to be rolled over by the rollover determination unit;
An image processing apparatus characterized in that the lung field search processing means searches for a lung field on the radiation image after rotation processing. In the image processing apparatus according to claim 1,
The lung field search processing means executes a lung field search on the radiation image which is determined by the rollover judging means that the subject image is vertically reflected in the radiation image. Image processing device. In the image processing apparatus according to claim 2,
Variation of pixel values in the area on the right side of the outline extraction image generated by extracting the outline of the image reflected on the radiation image determined to be lying down by the rollover judging means, and pixels in the area on the left side Comparing with the variation in value, it comprises a left / right determination means for determining that the region on which the variation is larger corresponds to the upper part of the subject image,
An image processing apparatus, wherein the image rotation unit rotates the radiographic image according to a determination result of the left / right determination unit so that an upper portion of the subject image is directed upward of the radiographic image. In the image processing apparatus according to claim 2,
In the upper region of the contour extraction image generated by extracting the contour of the image reflected in the radiation image in which the object image is determined to be reflected vertically in the radiation image by the rollover judging means An upper / lower determination unit configured to compare the variation of the pixel value with the variation of the pixel value in the lower region and determine that the region having the larger variation corresponds to the upper portion of the object image;
An image processing apparatus, wherein the image rotation unit rotates the radiographic image so that an upper portion of a subject image positioned on the lower side of the radiographic image faces an upper side of the radiographic image.   A program causing a computer to function as each unit of the image processing apparatus according to any one of claims 1 to 5. A radiation imaging apparatus comprising the image processing apparatus according to any one of claims 1 to 5.

Images