JP4891956B2 - Image processing device - Google Patents

Description

  The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus that classifies fat.

  Fat is roughly classified into subcutaneous fat formed under the skin and visceral fat formed around the viscera. Of these, visceral fat is attracting attention in the diagnosis of lifestyle-related diseases and the like. Therefore, it is desirable not only to measure the amount of body fat that reflects the total amount of fat, but also to determine the type of fat and measure the amount of fat for each type.

  Under such circumstances, the inventor of the present application proposes a technique for classifying fat in a tomographic image obtained from an X-ray CT apparatus or the like in Patent Document 1. Patent Document 1 describes an epoch-making technique capable of discriminating subcutaneous fat and visceral fat in, for example, a tomographic image of an X-ray CT apparatus.

JP 2003-339694 A

  The inventor of the present application has conducted research and development on a revolutionary technology improvement proposed in Patent Document 1. In particular, research and development has been repeated with the goal of more accurate fat types.

  The present invention has been made in the course of research and development, and an object thereof is to classify fat with high accuracy in a tomographic image.

  In order to achieve the above object, an image processing apparatus according to a preferred aspect of the present invention includes a pixel identification unit that identifies a muscle pixel and a fat pixel based on a pixel value of each pixel in a tomographic image of a subject, Moving the subcutaneous fat cursor from the outside to the inside of the subject excluding the skin pixel in the tomographic image, the skin determination unit for determining the muscle pixel constituting the outline of the subject in the tomographic image as a skin pixel, And a fat type section that determines a fat pixel that has passed until the subcutaneous fat cursor reaches the muscle pixel as a subcutaneous fat pixel.

  In a preferred aspect, the fat type section determines fat pixels excluding subcutaneous fat pixels in the tomographic image as visceral fat pixels.

  In a preferred aspect, the fat type section moves a subcutaneous fat cursor toward the center of gravity of the subject excluding skin pixels in the tomographic image.

  In a preferred aspect, the fat type section moves the subcutaneous fat cursor for each straight line along a plurality of straight lines radially extending around the center of gravity position.

  In a preferred embodiment, the subcutaneous fat cursor has a width in a direction perpendicular to the moving direction larger than a discontinuity width of muscle pixels excluding skin pixels.

  In order to achieve the above object, a program according to a preferred aspect of the present invention includes a pixel identification function for identifying a muscle pixel and a fat pixel based on a pixel value of each pixel in a tomographic image of the subject, A skin determination function for determining the muscle pixels constituting the outline of the subject in the tomographic image as skin pixels, and moving the subcutaneous fat cursor from the outside to the inside of the subject excluding the skin pixels in the tomographic image, A fat type function for determining, as a subcutaneous fat pixel, a fat pixel that has passed until the subcutaneous fat cursor reaches the muscle pixel is realized by the computer.

  For example, the program of the above aspect is stored in a storage medium such as a disk or a memory, and is read into the computer via the storage medium. Alternatively, the program may be provided to the computer via a network or the like.

  According to the present invention, fat can be classified with high accuracy in a tomographic image. For example, according to a preferred aspect of the present invention, it is possible to identify subcutaneous fat and visceral fat with high accuracy in a tomographic image.

  FIG. 1 is a diagram for explaining a preferred embodiment of the present invention. FIG. 1 shows an image processing system using an image processing apparatus according to the present invention. The system shown in FIG. 1 includes an X-ray CT apparatus 10 and an image processing apparatus 30. In the present embodiment, the X-ray CT apparatus 10 shown in FIG. 1 is used, but another known X-ray CT apparatus may be used instead of the X-ray CT apparatus 10 shown in FIG. Furthermore, instead of the X-ray CT apparatus, another image forming apparatus may be used.

  The X-ray CT apparatus 10 has a dynamic variable function of magnification. For example, an animal such as a mouse is a measurement target. Of course, other specimens may be measured. The X-ray CT apparatus 10 is provided with an X-ray generator 52 on one side and an X-ray detector 60 on the other side with the rotation center axis O therebetween. A collimator 54 is provided on the irradiation side of the X-ray generator 52. The X-ray generator 52 generates a divergent or fan-shaped (here, fan-beam shaped) X-ray beam 56 as shown. On the other hand, the X-ray detector 60 is configured as a plurality of (for example, 100) X-ray sensors arranged in a line, and an X-ray receiving opening is set according to the opening angle of the X-ray beam 56. Incidentally, the arrangement of the plurality of X-ray sensors may be linear or arcuate.

  In FIG. 1, a high voltage source used with the X-ray generator 52, a data processing circuit used with the X-ray detector 60, and the like are not shown.

  In FIG. 1, reference numeral 58 denotes an effective field of view. This is a circular region in which a CT image can be formed when the X-ray beam 56 is rotationally scanned. Incidentally, the effective visual field 58 is determined according to the positional relationship between the subject or the rotation center axis O and the X-ray generator 52 and the X-ray detector 60. In the present embodiment, since a displacement mechanism 62 described below is provided, it is possible to change the positional relationship between them and mechanically vary the magnification of the CT image.

  In other words, the X-ray generator 52 and the X-ray detector 60 are connected to the displacement mechanism 62. In this embodiment, the displacement mechanism 62 is a distance between the X-ray generator 52 and the X-ray detector 60. The function of displacing them (that is, the measurement unit) in the beam axis direction of the X-ray beam 56 is maintained. In this case, the rotation center axis O is not changed, that is, the magnification can be changed by moving the measurement unit without moving the subject. The displacement mechanism 62 includes a motor 62A for generating a displacement force. In the present embodiment, the displacement amount can be individually set for each CT imaging, that is, the measurement magnification can be individually set. When such control is performed, the shape of the housing member (especially the outer diameter at the CT imaging position) is used as a reference, and the housing member is prevented from contacting or colliding with the gantry or the measurement unit when the magnification is variable. Has been.

  The gantry rotation mechanism 66 is a mechanism that rotates the entire base including the displacement mechanism mounted thereon by rotating the rotation base. Since the measurement unit is mounted on the displacement mechanism 62, the measurement unit positioned at a desired position by the displacement mechanism 62 is rotationally driven while maintaining the position. The gantry rotating mechanism 66 has a motor 66A for generating the driving force. Note that the measurement magnification does not change during one rotation. The slide mechanism 68 is a moving mechanism that slides an arm (not shown), and the driving force is generated by a motor 68A. The arm slides the mounting table on which the subject is placed in the direction of the rotation center axis. Then, CT imaging of the subject is performed at each slide position.

  Incidentally, although various mechanisms 62, 66, 68 and the like are shown in FIG. 1, it is desirable to provide a sensor in order to detect a position or a change due to these mechanisms. Then, it is desirable that an arithmetic control unit (not shown) performs so-called feedback control based on the output signals of those sensors. In the example shown in FIG. 1, the slide mechanism 68 has the motor 68A as a drive source. However, the slide force may be artificially generated. The overall operation of the X-ray CT apparatus 10 is controlled by an operation control unit (not shown). The operation control unit is preferably provided inside the X-ray CT apparatus 10.

  The image processing apparatus 30 is configured by hardware such as a CPU, a memory, and a hard disk, and software such as an operation control program. FIG. 1 shows typical functions realized by the cooperation of the hardware and software. That is, the image processing apparatus 30 includes a pixel identification unit 310, a skin determination unit 320, a fat type unit 330, a fat amount calculation unit 340, and the like. The computer may be operated as the image processing apparatus 30 using a program corresponding to these functions.

  Note that a display device 32, a storage device 34, a keyboard 36, a mouse 38, a printer 40, and the like are connected to the image processing device 30. The storage device 34 stores X-ray CT tomographic image data 34A as necessary.

  In the present embodiment, the tomographic image obtained by the X-ray CT apparatus 10 is processed by the image processing apparatus 30, and subcutaneous fat and visceral fat are identified with high accuracy in the tomographic image. Therefore, each function in the image processing apparatus 30 will be described in detail below. In addition, about the part (structure) shown in FIG. 1, the code | symbol of FIG. 1 is utilized also in the following description.

  FIG. 2 is a diagram illustrating an example of an image processed in the image processing apparatus 30. A tomographic image 100 shown in FIG. 2 is a schematic view of an X-ray CT image obtained by imaging the abdomen of a subject (eg, an animal such as a mouse) with the X-ray CT apparatus 10.

  In the tomographic image 100, in order from the outside of the subject, skin pixels 112 corresponding to skin, subcutaneous fat pixels 122 corresponding to subcutaneous fat, abdominal muscle pixels 114 corresponding to abdominal muscles, visceral fat pixels 124 corresponding to visceral fat, The internal pixel 116 corresponding to internal organs, such as an intestine, is included. In the tomographic image 100, a background pixel 130 corresponding to air or the like exists around the subject.

  When the tomographic image 100 is formed by the X-ray CT apparatus 10, image data corresponding to the tomographic image 100 is supplied to the image processing apparatus 30. First, the image processing apparatus 30 executes a pixel identification process in the tomographic image 100. That is, the pixel identification unit 310 identifies muscle pixels, fat pixels, and other pixels based on the pixel value (CT value) of each pixel in the tomographic image 100.

  FIG. 3 is a diagram illustrating a tomographic image 100 in which pixels are identified. When attention is paid to each pixel in the tomographic image 100 obtained by the X-ray CT apparatus 10, a CT value corresponding to the degree of X-ray absorption in the pixel portion corresponds. The CT value of water is defined as 0 and the CT value of air is defined as -1000. For example, the CT value of fleshy tissue such as muscle is about -120 to +350, and the CT value of fat is about -500 to -180. .

  Therefore, the pixel identifying unit 310 compares all the pixels in the tomographic image 100 with a pixel value (CT value) of each pixel and a threshold value, as shown in FIG. The pixel 110 is divided into a fat pixel 120 corresponding to fat and a background pixel 130 corresponding to air.

  It should be noted that in the stage of FIG. 3 subjected to the identification processing by the pixel identification unit 310, the subcutaneous fat formed under the skin and the visceral fat formed around the viscera are not discriminated and are both identified as the fat pixel 120. Yes.

  Next, the skin determination unit 320 determines that the muscle pixel 110 constituting the outline of the subject is a skin pixel in the tomographic image 100 of FIG. The skin determination unit 320 uses the boundary portion with the background pixel 130 as the outermost contour of the subject, and deletes only the muscle pixels 110 by a predetermined number of pixels (for example, 10 pixels) from the outermost contour toward the inside of the subject (fat) Pixel 120 is not deleted). Thus, the muscle pixel 110 deleted by the skin determination unit 320 is regarded as a skin pixel.

  FIG. 4 is a diagram illustrating an image of the subject from which the skin pixel 112 has been identified. By the process in the skin determination unit 320, the skin pixel 112 that is a muscle pixel constituting the outline of the subject is identified, and only the skin pixel 112 is deleted from the image of the subject. When the skin pixel 112 is deleted, the fat type section 330 identifies subcutaneous fat and visceral fat.

  FIG. 5 is a diagram for explaining processing in the fat type section 330. The fat type section 330 calculates the center of gravity of the subject from which the skin pixels have been removed, and sets the center of gravity position P. Then, the fat type section 330 moves the subcutaneous fat cursor 200 from the outside of the subject excluding the skin pixel toward the center of gravity P of the subject, and the fat that has passed until the subcutaneous fat cursor 200 reaches the muscle pixel 110. The pixel 120 is determined as a subcutaneous fat pixel.

  The subcutaneous fat cursor 200 moves along a straight line from the outside of the subject toward the center of gravity P. That is, as shown in FIG. 5, the subcutaneous fat cursor 200 moves from the outside of the subject toward the center of gravity P along the dashed arrow. The subcutaneous fat cursor 200 is moved until it reaches the muscle pixel 110. That is, the subcutaneous fat cursor 200 moves until reaching the muscle pixel 110 corresponding to the abdominal muscle pixel (reference numeral 114 in FIG. 2).

  For example, it is determined that the mouse has reached the position where the subcutaneous fat cursor 200 hits the muscle pixel 110 (the position where the muscle pixel 110 enters the subcutaneous fat cursor 200). Alternatively, it may be determined that the subcutaneous fat cursor 200 has reached the muscle pixel 110 and has reached N pixels further. The number of pixels N at that time is empirically set to N = 2, for example.

  Then, the fat type section 330 determines the subcutaneous fat pixel by moving the subcutaneous fat cursor 200 along the straight line from the omnidirectional surrounding the subject toward the gravity center position P. For example, a plurality of straight lines are extended radially around the center of gravity position P, and the subcutaneous fat cursor 200 is moved from the outside of the subject toward the center of gravity position P for each straight line to determine subcutaneous fat pixels. Thus, subcutaneous fat pixels are determined in all directions (all angles) surrounding the subject, and subcutaneous fat pixels existing so as to surround the subject excluding skin pixels are identified.

  Note that the abdominal muscle pixels (reference numeral 114 in FIG. 2) may not form a closed curve in the tomographic image. For example, as shown in FIG. 5, the muscle pixel 110 corresponding to the abdominal muscle pixel may be interrupted on the upper side of the drawing. Therefore, it is desirable that the width of the subcutaneous fat cursor 200 in the direction perpendicular to the moving direction is larger than the maximum discontinuity width of the muscle pixel 110. For example, the width of the subcutaneous fat cursor 200 is set to about 10 pixel width. As a result, the subcutaneous fat cursor 200 reliably reaches the muscle pixel 110 even in the interrupted portion of the muscle pixel 110, and the subcutaneous fat pixel can be identified even in the interrupted portion of the muscle pixel 110.

  Incidentally, in FIG. 5, the subcutaneous fat cursor 200 has an arc shape in accordance with the shape of the abdominal muscles, but the shape of the subcutaneous fat cursor 200 is not limited to the example of FIG. 5.

  FIG. 6 is a diagram illustrating an image of a subject in which the subcutaneous fat pixels 122 are identified. The fat type section 330 determines the subcutaneous fat pixel 122 using a subcutaneous fat cursor (reference numeral 200 in FIG. 5). The remaining fat pixels excluding the subcutaneous fat pixels 122 are determined as visceral fat pixels 124. In this way, the subcutaneous fat formed under the skin and the visceral fat formed around the viscera are discriminated, and the subcutaneous fat and the visceral fat are identified with high accuracy in the tomographic image.

  In the subject image shown in FIG. 6, the muscle pixel surrounded by the visceral fat pixels 124 may be determined as the visceral pixel 116 corresponding to the visceral organ such as the intestine, and further, the visceral pixel 116 is excluded. The remaining muscle pixels may be determined as the abdominal muscle pixels 114.

  When the subcutaneous fat pixel 122 and the visceral fat pixel 124 are discriminated by the fat classification unit 330, the fat amount calculation unit 340 calculates the fat amount of the visceral fat based on the number of the visceral fat pixels 124 included in the tomographic image. To do. For example, by multiplying the area per pixel by the number of visceral fat pixels 124, the area of visceral fat is calculated as an index indicating fat mass. Of course, the area of visceral fat relating to a plurality of tomographic images may be calculated, and the volume of visceral fat may be calculated as an index indicating the amount of fat. Furthermore, the amount of subcutaneous fat may be calculated, and the amount of abdominal muscles may be calculated as necessary.

  As mentioned above, although preferred embodiment of this invention was described, according to embodiment mentioned above, even if it is a case where an abdominal muscle is interrupted, it becomes possible to recognize a visceral fat automatically comparatively correctly. Therefore, the processing time for recognition can be shortened compared with the case where visceral fat is recognized by adding manual correction by an operator, for example. Further, according to the above-described embodiment, since visceral fat is automatically recognized according to a certain rule, objective and highly reproducible measurement is possible. The above-described embodiments are merely examples in all respects, and do not limit the scope of the present invention.

It is a figure which shows the image processing system using the image processing apparatus which concerns on this invention. It is a figure which shows the example of an image processed in an image processing apparatus. It is a figure which shows the tomographic image in which the pixel was identified. It is a figure which shows the image of the subject by which the skin pixel was identified. It is a figure for demonstrating the process in a fat classification part. It is a figure which shows the image of the subject by which the subcutaneous fat pixel was identified.

Explanation of symbols

  30 image processing device, 310 pixel identification unit, 320 skin determination unit, 330 fat type unit, 340 fat amount calculation unit.

Claims (6)

A pixel identifying unit that identifies muscle pixels and fat pixels based on the pixel value of each pixel in the tomographic image of the subject;
A skin determination unit that determines a muscle pixel constituting a contour of the subject in the tomographic image as a skin pixel;
Fat type that moves the subcutaneous fat cursor from the outside to the inside of the subject excluding the skin pixel in the tomographic image and determines that the fat pixel that has passed until the subcutaneous fat cursor reaches the muscle pixel is the subcutaneous fat pixel And
Having
An image processing apparatus. The image processing apparatus according to claim 1.
The fat type part determines fat pixels excluding subcutaneous fat pixels in the tomographic image as visceral fat pixels,
An image processing apparatus. The image processing apparatus according to claim 1 or 2,
The fat type part moves a subcutaneous fat cursor toward the center of gravity of the subject excluding skin pixels in the tomographic image,
An image processing apparatus. The image processing apparatus according to claim 3.
The fat type section moves a subcutaneous fat cursor for each straight line along a plurality of straight lines that are radially extended around the center of gravity position.
An image processing apparatus. In the image processing device according to any one of claims 1 to 4,
The subcutaneous fat cursor has a width in a direction perpendicular to the moving direction larger than the discontinuity width of muscle pixels excluding skin pixels,
An image processing apparatus. A pixel identification function for identifying muscle pixels and fat pixels based on the pixel value of each pixel in the tomographic image of the subject;
A skin determination function for determining a muscle pixel constituting a contour of the subject in the tomographic image as a skin pixel;
Fat type that moves the subcutaneous fat cursor from the outside to the inside of the subject excluding the skin pixel in the tomographic image and determines that the fat pixel that has passed until the subcutaneous fat cursor reaches the muscle pixel is the subcutaneous fat pixel Function and
A program characterized by causing a computer to realize.

Images