JP4626984B2 - Difference image generating apparatus and method - Google Patents

Description

  The present invention relates to a difference image generation apparatus, and more particularly to a difference image generation apparatus that generates a difference image with fewer false images when generating a time-dependent difference image.

Patent Document 1 discloses a differential image generation apparatus that generates a differential image by subtracting a past image and a current image. Specifically, in order to detect a shape change between the input image In1 (x, y) and the input image In2 (x, y), in the method of obtaining the difference image S (x, y), A reference area regarded as having no shape change is set, and average values ave1 and ave2 of the pixel values of the reference area are obtained, and a difference image S (x, y) is obtained as S (x, y) = In1 (x, y ) -In2 (x, y)-(ave1-ave2) and display this.
Japanese Patent Laid-Open No. 2002-190010

  The differential image generation apparatus of Patent Literature 1 determines an area where the amount of change in pixel value is small between a past image and a current image as a reference area. Therefore, in order to set the reference area, it is necessary to perform a difference process by aligning the past image and the current image. After that, the pixel value of the past image is corrected using the reference area, and the difference process is performed again between the corrected past image and the current image, so that the difference process needs to be performed at least twice.

  In addition, when the alignment process and the difference process for setting the reference area are not performed, a part where the fluctuation due to respiration is small, for example, the vicinity of the mediastinum is set as the reference area as a part where the reference area is set in advance. It is necessary to input information to the difference image generation apparatus. In this case, the reference region is limited to one location near the mediastinum as a region where the amount of change in pixel value due to respiration is small. When the reference area is set only in the vicinity of the mediastinum, since the pixel values are high in the vicinity of the mediastinum, correction of the density value cannot be sufficiently performed in other parts where the pixel values are low, such as lung field areas and rib areas. That is, since the contrast is different between the lung field region and the rib region and the mediastinum region, the rib region and the lung field region may not be corrected using the average value near the mediastinum region. Furthermore, in the differential image generating apparatus of Patent Document 1 in which a reference region is set in advance, individual differences between different subjects, secular changes such as fat condition in the same subject, breathing and posture at the time of imaging ( Necessary density correction cannot be performed for the displacement of the pixel value due to the difference in the degree of adhesion to the apparatus. As a result, a difference value may be output even in a part that does not change with time, and a false image may occur in the difference image.

  Accordingly, an object of the present invention is to provide a differential image generation apparatus that is unlikely to output a differential value at a portion that does not change with time, and that can prevent generation of a false image.

  In order to achieve the above object, a difference image generating apparatus according to the present invention includes a first medical image and a second medical image obtained by capturing the same part of a subject at different dates and times, For the medical image, a local region is set for each region photographed in the first medical image, and a local region setting means for setting a local region in a region corresponding to the region for the second medical image; Pixel value calculating means for calculating a pixel value representative of the local region included in the first medical image and a pixel value representative of the local region included in the second medical image; and the pixel value calculating means included in the second medical image. A correction for correcting a pixel value representing the local area in the first medical image so that a pixel value representing the local area of the first medical image corresponding to the local area approaches a pixel value representing the local area. And means, wherein the difference image generation means for generating a difference image by subtracting the corrected said the first medical image second medical image by correcting means, and display means for displaying the difference image.

  The pixel value calculating means calculates a maximum pixel value under a predetermined condition among pixels constituting the local region included in the first medical image and the second medical image, and the correcting means Calculating a pixel value correction curve based on a difference between the maximum pixel value of the local region included in one medical image and the maximum pixel value of the local region included in the second medical image; Based on the above, the pixel value of the local region of the first medical image is corrected.

  The “maximum pixel value under a predetermined condition” as used herein refers to the maximum pixel value after performing noise removal processing on the image data, or the predetermined pixel value in the image data after noise removal. The pixel value of a pixel that satisfies the condition that it is regarded as the maximum pixel value only when there are a predetermined number or more of pixels having.

  In addition, the medical image acquisition unit acquires a first medical image and a second medical image obtained by imaging lung field regions of the same subject taken at different dates and times, and the local region setting unit includes the first medical image and A rib area included in the second medical image is extracted, and the local area is set along the rib area.

  In addition, the medical image acquisition unit acquires a first medical image and a second medical image obtained by imaging lung field regions of the same subject taken at different dates and times, and the local region setting unit includes the first medical image and A rectangular local region is set for the second medical image, or left and right lung field regions included in the first medical image and the second medical image are extracted, the left lung region is divided into two, and the right lung A region is divided into three to set a local region, and a rib region and a lung field region of the subject are extracted from the local region, and the correction unit is configured to generate a pixel value for each of the extracted rib region and lung field region. Perform the correction.

  In addition, a medical image acquisition step for obtaining a first medical image and a second medical image obtained by photographing the same part of the subject at different dates, and for each part photographed in the first medical image with respect to the first medical image A local region setting step for setting a local region and setting a local region in a region corresponding to the part with respect to the second medical image, a pixel value representing the local region included in the first medical image, and the second medical image A pixel value calculating step for calculating a pixel value representative of the local area included in the second medical image; and a pixel value representative of the local area included in the second medical image; A correction step of correcting a pixel value representative of the local region in the first medical image so that a pixel value representative of the local region of one medical image approaches, and the first medical image corrected by the correction step; A difference image generation step of generating a difference image and a serial second medical image and differences, including a display step of displaying the difference image may be a difference image generation method.

  According to the present invention, since a difference image between a medical image whose density value has been corrected for each local region and another medical image is generated, the difference value is less likely to be output at a portion where there is no change over time, and more false images are generated. It is possible to provide a differential image generation apparatus that can prevent the above-described problem.

  Hereinafter, embodiments of a differential image generating apparatus according to the present invention will be described with reference to the accompanying drawings. Note that components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiment of the invention, and the repetitive description thereof is omitted.

  The difference image generation apparatus according to the present embodiment reads a past image and a current image obtained by photographing the same part of the same subject at intervals, and generates a difference image by performing a difference process between the past image and the current image. . This difference image depicts a portion of the same portion of the same subject that has changed over time.

[First embodiment]
FIG. 1 is a hardware configuration diagram showing a configuration of a difference image generation apparatus according to an embodiment of the invention.

  The differential image generation apparatus 10 in FIG. 1 stores medical image imaging apparatuses 1 such as an X-ray CT apparatus, an MR apparatus, an X-ray apparatus, and a DR apparatus, and an image that stores medical images obtained by the medical image imaging apparatus 1. A database 2 and a difference image generation apparatus 10 that generates and displays a difference image by subtracting two medical images, and includes a medical image photographing apparatus 1, an image database (image DB) 2, and a difference image generation apparatus 10 are connected to each other by a network such as LAN3.

  The difference image generation apparatus 10 mainly includes a central processing unit (CPU) 11 that controls the operation of each component, a main memory 12 that stores a control program for the apparatus and that serves as a work area when the program is executed, and an operating system. A system (OS), a device drive of a peripheral device, a magnetic disk 13 storing various application software including a program for generating and displaying a difference image by differentially processing two medical images, and display data A display memory 14 for temporary storage, a monitor 15 such as a CRT monitor or a liquid crystal monitor for displaying an image based on data from the display memory 14, a keyboard 16, a mouse 17 as a position coordinate input device, and a mouse 17 Signals such as the position of the mouse pointer on the monitor 15 and the state of the mouse 17 are detected. It includes a controller 18 for outputting to PU11, and a common bus 19 for connecting the above components.

  In the present embodiment, the differential image generation apparatus 10 reads a medical image from the image database 2 via the LAN 3, but a storage device connected to the differential image generation apparatus 10, for example, an FDD, a CD-RW drive, an MO drive A medical image may be read from a ZIP drive or the like. Alternatively, a medical image may be acquired directly from the medical image photographing apparatus 1 via the LAN 3.

  Next, a procedure for generating and displaying a difference image using the difference image generating apparatus 10 according to the present embodiment will be described with reference to FIG. FIG. 2 is a flowchart showing a procedure of difference image generation and display processing. In the following, description will be given in the order of each step in FIG.

(Step S1)
The difference image generation device 10 acquires a past image and a current image in which the same part of the same subject is copied from the medical image capturing device 1 or the image database 2. Then, a lung field region is extracted from the past image and the current image by outline extraction processing or the like (S1).

(Step S2)
Using the current image as a reference, the past image is subjected to parallel movement processing, rotation processing, enlargement / reduction processing, and alignment processing between the current image and the past image (hereinafter referred to as “alignment image”) is performed (S2). The alignment process is performed by matching the contour lines of the current image and the alignment image using the contour information of the current image and the alignment image extracted in step S1.

(Step S3)
In step S2, a local region is set for the alignment image and the current image that have undergone parallel movement processing, rotation processing, and the like for alignment processing. Then, the pixel value (hereinafter referred to as “density value”) of the pixels constituting the local area set for the alignment image and the density value of the corresponding local area set for the current image are close to each other. Processing for correcting the density value of the local region of the alignment image is performed (S3). Details of the local density value correction processing will be described later.

(Step S4)
A difference image is generated by performing a difference process between the alignment image in which the local density value has been corrected in step S3 and the current image (S4). Then, the difference image is displayed on the monitor 15.

  Next, local region correction processing will be described with reference to FIGS. FIG. 3 is a schematic diagram showing the content of the local density value correction process. FIG. 4 is a flowchart showing a flow of local density value correction processing. FIG. 5 is a schematic diagram illustrating an example of setting a local region. FIG. 6A is a diagram in which the difference between the average maximum density values in the local region is plotted along the body axis direction, and FIG. 6B is a diagram illustrating an example of the density value correction curve. 7 and 8 are schematic diagrams illustrating other setting examples of the local region.

In FIG. 3A, the density profile of the current image is indicated by a solid line, and the density profile of the alignment image is indicated by a dotted line. The density profile is obtained by plotting the pixels constituting the medical image on the corresponding coordinates, with the pixel value on the horizontal axis and the number of pixels on the vertical axis.

  Since the alignment image and the current image are obtained by photographing the same part of the same subject, the density profiles should substantially match. However, the difference in imaging conditions at the time of imaging (for example, in the case of an axial image, the degree of adhesion when the subject is brought into close contact with the X-ray imaging apparatus, or the shape of the subject when the past image and the current image are taken) The difference value may also occur in an area that originally does not change with time and the density value of the pixel should be unchanged. This appears as a difference between the peak value (mode) of the number of pixels and the maximum value of the density value in the density profile of the current image and the density profile of the alignment image in FIG.

  Therefore, by performing the local area density correction processing on the alignment image, as shown in FIG. 3B, the baseline of the alignment image is set so that the density profile of the alignment image approaches the density profile of the current image. The process of raising or lowering is performed. Hereinafter, the local region correction processing will be described along each step of FIG.

(Step S3)
The local area density value correction process is started.

(Step S31)
A local region is set for the alignment image and the current image (S31). In the local region setting process, for example, as shown in FIG. 5, the rib region may be extracted from each of the alignment image and the current image, and the region may be set along the rib. In FIG. 5, local regions R1, R2,..., R13 are set in the right lung region, and L1, L2,. In FIG. 5, local regions with even subscripts (R2, L2, R4, L4,..., R12, L12) are rib regions, and odd local regions (R1, L1, R3, L3,... R13, L13) are lungs. Indicates the field. Further, as an example of setting the local region, as shown in FIG. 7, geometrically rectangular local regions R71, R72,... R715, L71, L72,. In this case, the rib region and the lung field region are extracted for each of the rectangular local regions R71, R71,... R75, L71, L72,. Then, the rib area and the lung field area are extracted from each local area, and the density values of the rib area and the lung field area of each local area are corrected by the following processing. Further, as shown in FIG. 8, the left lung field region may be divided into two and the right lung field region may be divided into three along the anatomical viewpoint, and each divided region may be set as a local region. Also in this case, the rib region and the lung field region are extracted for each of the upper left local region L81, the lower left local region L82, the upper right local region R81, the right middle local region R82, and the lower right local region R83. The density value of the field area is corrected by the following processing.

(Step S32)
The average maximum density value of each local area set in step S31 is calculated (S32). Here, the “average maximum density value” refers to an average value of pixel values of n pixels, in which n pixels are extracted in descending order of pixel values from among pixels constituting the local region.

Where P: the average maximum density value of the local region
P _i : The pixel value of the i-th pixel in order from the pixel corresponding to the maximum pixel value among the pixels constituting each local region. Note that threshold processing is performed when n pixels are counted in order to remove the influence of noise. The pixel is counted as a pixel for calculating the average maximum density value only when there are a predetermined number of pixels having a certain density value, for example, 10 pixels or more. In Equation 1, there are 10 or more pixels corresponding to the density value P.

(Step S33)
The difference between the average maximum density values of the local areas of the alignment image and the current image calculated in step S32 is rearranged along the body axis direction. Then, a density correction curve that is a curve for interpolating each point is generated. For the generation of the interpolation curve, a linear function, a nonlinear function, a least square method, a spline curve, or the like is appropriately used. Using the density correction curve, the density value of the alignment image is corrected according to Equation 2 (S33).

[Equation 2]
P ′ (x, y) = P (x, y) + D (y)
P (x, y): Pixel value of the alignment image before correction processing P ′ (x, y): Pixel value of the alignment image after correction processing
D (y): Density Value Correction Curve FIG. 6 shows an example of the density correction curve. The horizontal axis indicates the position in the body axis direction (distance in the direction from the head toward the foot), and the vertical axis indicates the average maximum pixel value corresponding to each position. Each point in FIG. 6 plots the difference between the position of each local region (for example, R1, R2,... R13) in the body axis direction and the average maximum pixel value of the corresponding alignment image and the average maximum pixel value of the current image. This is the point. A curve for interpolating each point is generated, and a density correction curve indicated by a dotted line is calculated. The density correction curve is added to the density value of the alignment image to smoothly correct the density value of a continuous local region. Thereby, as shown in FIG. 3B, the baseline of the density profile can be raised or lowered for each local region of the alignment image, and the baseline of the alignment image and the current image can be aligned.

  The density value correction may be performed using the average maximum pixel value calculated in step S32 without generating the density value correction curve. However, if the density value correction is performed using the density value correction curve, a continuous area is displayed. On the other hand, smoother density value correction can be performed.

  According to the difference image generation apparatus according to the present embodiment, the pixel value of the alignment image (past image) is corrected for the density value for each local region along the body axis direction of the subject, and then the difference from the current image. To do. As a result, differences in imaging conditions between the alignment image and the current image, secular changes in the subject, and differences in density values due to posture devices at the time of imaging can be corrected for each region of the subject. A part that does not change with time is not easily reflected by the difference image, and a false image is less likely to occur. As a result, it is possible to clearly depict only a portion having a change with time.

Hardware configuration diagram showing the configuration of the difference image generation device The flowchart which shows the process sequence of the difference image generation method Schematic diagram showing the contents of local area correction processing Flowchart showing the flow of local area correction processing Schematic diagram showing an example of setting a local region FIG. 6A is a schematic diagram illustrating a generation process of a density value correction curve, in which FIG. 6A is a diagram in which a difference between average maximum density values in a local region is plotted along a body axis direction, and FIG. Diagram showing an example of a correction curve Schematic diagram showing another setting example of the local region Schematic diagram showing another setting example of the local region

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Medical imaging device, 2 ... Image database, 3 ... LAN, 10 ... Difference image generation apparatus, 11 ... CPU, 12 ... Main memory, 13 ... Magnetic disk, 14 ... Display memory, 15 ... Monitor, 16 ... Keyboard, 17 ... Mouse, 18 ... Controller, 19 ... Common bus

Claims (4)

Medical image acquisition means for acquiring a first medical image and a second medical image obtained by imaging the same part of the subject at different dates and times;
A local area setting unit that sets a local area for each part imaged in the first medical image for the first medical image and sets a local area in an area corresponding to the part for the second medical image When,
Pixel value calculating means for calculating a pixel value representative of the local region included in the first medical image and a pixel value representative of the local region included in the second medical image;
A difference between a pixel value representative of the local area included in the second medical image and a pixel value representative of the local area of the first medical image corresponding to the local area; and a position of the local area; Based on the difference between the representative pixel values at each position, a correction curve that determines a correction value of the pixel value corresponding to the pixel position is calculated, and the pixel of the pixel to be corrected is calculated using the correction curve. Correction means for correcting the pixel value by obtaining a correction value of the pixel value corresponding to the position and adding the correction value to the pixel value of the first medical image ;
Difference image generation means for generating a difference image by subtracting the first medical image and the second medical image corrected by the correction means;
Display means for displaying the difference image;
A difference image generation apparatus comprising: The pixel value calculating means extracts a predetermined number of pixels in order from the pixel value having the largest pixel value among the pixels constituting the local region included in the first medical image and the second medical image, and the predetermined number Calculate the average maximum pixel value, which is the average of the pixel values of the pixels,
The difference image generation apparatus according to claim 1, wherein the correction unit uses the average maximum pixel value as a pixel value representing the local region . The pixel value calculation unit extracts the predetermined maximum number of pixels as a pixel for calculating the average maximum pixel value only when there are a predetermined number or more of pixels having the same pixel value. To
The difference image generation apparatus according to claim 2 , wherein A medical image acquisition step of acquiring a first medical image and a second medical image obtained by imaging the same part of the subject at different dates and times;
A local region setting step of setting a local region for each part photographed in the first medical image for the first medical image and setting a local region in a region corresponding to the part for the second medical image When,
A pixel value calculating step for calculating a pixel value representative of the local region included in the first medical image and a pixel value representative of the local region included in the second medical image;
A difference between a pixel value representative of the local area included in the second medical image and a pixel value representative of the local area of the first medical image corresponding to the local area; and a position of the local area; Based on the difference between the representative pixel values at each position, a correction curve that determines a correction value of the pixel value corresponding to the pixel position is calculated, and the pixel of the pixel to be corrected is calculated using the correction curve. A correction step of correcting the pixel value by obtaining a correction value of the pixel value corresponding to the position and adding the correction value to the pixel value of the first medical image;
A difference image generation step of generating a difference image by subtracting the first medical image and the second medical image corrected by the correction step;
A display step of displaying the difference image;
The difference image generation method characterized by including.

Images