JP6334942B2 - Medical image processing apparatus, medical image processing method, and program - Google Patents

Description

  Embodiments described herein relate generally to a medical image processing apparatus, a medical image processing method, and a program for processing image data obtained using an image diagnostic apparatus such as an X-ray CT apparatus and an MRI apparatus.

  There is an image filter for image data obtained using an image diagnostic apparatus such as an X-ray CT (Computed Tomography) apparatus or an MRI (Magnetic Resonance Imaging) apparatus. This image filter is used particularly when the bronchial region is emphasized on the image data. For example, in bronchial diagnosis, bronchial wall thickness and stenosis may be quantitatively diagnosed, but using the image after this image filter is expected to improve the segmentation accuracy of the bronchial lumen and bronchial wall. it can.

  As a conventional method, there is a method using a general enhancement filter. Specifically, first, an edge enhancement filter such as a Sobel filter is applied. Thereafter, an image in which the vicinity of the image contour is emphasized is created by adding the filter image to the original image.

JP 2010-136765 A

  In the above method, since the entire image is filtered, it takes a long time. Also, since anatomical information is not taken into account, it tends to be vulnerable to noise.

  An object of the present invention is to provide a medical image processing apparatus, a medical image processing method, and a program that can provide more accurate diagnosis support in quantitative diagnosis of a bronchial region.

  According to the embodiment, the medical image processing apparatus includes a pulmonary artery extraction unit, an existence probability calculation unit, and an enhanced image generation unit. The pulmonary artery extraction unit extracts a pulmonary artery region from the medical image. The existence probability calculation unit calculates the existence probability of the bronchial region from the pulmonary artery region extracted by the pulmonary artery extraction unit. The emphasized image generation unit generates an image in which the bronchial region is emphasized based on the presence probability calculated by the presence probability calculation unit.

1 is a configuration diagram showing an example of a configuration of a medical image diagnostic system according to an embodiment. 5 is a flowchart illustrating a processing procedure of the medical image processing apparatus according to the present embodiment. The flowchart which shows the detailed process sequence of step ST2e of the said FIG. The figure which shows the image of the cross section orthogonal to the core line of a pulmonary artery area | region in this embodiment. The figure which shows the anatomical positional relationship of a pulmonary artery and a bronchus in this embodiment. The figure which shows the candidate for determining the center point of a bronchus in this embodiment. The figure which shows the definition of the processing range of a bronchial emphasis filter in this embodiment. The figure which shows the positional relationship of a bronchial wall and a bronchial lumen in this embodiment. The figure which shows the result of having performed the filter process about the pixel on a certain straight line in this embodiment. The figure which shows the color map image of the presence probability regarding a bronchial lumen in this embodiment. The figure which shows the image of the cross section orthogonal to the core line of the pulmonary artery area | region where the bronchus emphasis filter was given in this embodiment.

The figure which shows the color map image of the existence probability regarding a bronchial lumen.

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

  FIG. 1 is a configuration diagram illustrating an example of a configuration of a medical image diagnostic system according to the present embodiment. The medical image diagnostic system includes a medical image processing apparatus 100, a medical image photographing apparatus 200, and a medical image database 300. The medical image processing apparatus 100, the medical image photographing apparatus 200, and the medical image database 300 are electrically connected via a network (not shown). Note that wireless connection may be used instead of the network.

  The medical imaging apparatus 200 is, for example, an X-ray computed tomography (hereinafter referred to as CT) apparatus, a magnetic resonance imaging (hereinafter referred to as MRI) apparatus, or a nuclear medicine diagnostic apparatus. The medical image photographing apparatus 200 transfers the image data to the medical image database 300 via the network and stores it.

  The medical image processing apparatus 100 includes an input unit 110, a pulmonary artery extraction unit 120, an existence probability calculation unit 130, an image generation unit 140, a display unit 150, and a bronchial region extraction unit 160. The input unit 110 reads image data stored in the medical image database 300.

  The pulmonary artery extraction unit 120 extracts a pulmonary artery region from the image data read by the input unit 110. The existence probability calculating unit 130 calculates the existence probability of the bronchial region from the pulmonary artery region extracted by the pulmonary artery extracting unit 120. The image generation unit 140 generates an image in which the bronchial region is emphasized based on the presence probability calculated by the presence probability calculation unit 130.

  The display unit 150 displays the image generated by the image generation unit 140. The bronchial region extraction unit 160 extracts a bronchial region using the image generated by the image generation unit 140 and supplies the extracted bronchial region to the display unit 150. Then, the display unit 150 displays the bronchial region extracted by the bronchial region extracting unit 160 on the image data read from the medical image database 300 in a superimposed manner.

  Further, the image generation unit 140 generates a color map image based on the presence probability calculated by the presence probability calculation unit 130 and supplies the color map image to the display unit 150.

Next, the operation in the above configuration will be described.
FIG. 2 is a flowchart showing a processing procedure of the medical image processing apparatus 100. FIG. 3 is a flowchart showing a detailed processing procedure of the medical image processing apparatus 100.

  In this method, information on pulmonary arteries running along the bronchi is used to improve the accuracy of bronchi extraction. The pulmonary artery region has a contrast difference with the surroundings compared to the bronchial region, and is easily extracted. Details are described below.

  The medical image processing apparatus 100 reads the image data obtained by the medical image photographing apparatus 200 from the medical image database 300 (step ST2a). The medical image processing apparatus 100 extracts a pulmonary artery region from the data read in step ST2a (step ST2b), and an existing technique can be used for the method. For example, there is a region expansion method.

Subsequently, the medical image processing apparatus 100 detects the core line of the pulmonary artery region extracted in step ST2b (step ST2c), and an existing technique can be used for the method. An example is thinning. As shown in FIG. 4, the medical image processing apparatus 100 defines a cross section orthogonal to the core line detected in step ST2c (step ST2d), and an existing technique can be used for the method. Hereinafter, this cross section is referred to as a cross cut.
The medical image processing apparatus 100 applies a bronchus enhancement filter to the cross-cut section defined in step ST2d (step ST2e), details of which will be described later.

(Details of enhancement filter for bronchial region)
FIG. 3 shows a detailed processing flow of the enhancement filter for the bronchial region. In this method, as shown in FIG. 5, the anatomical positional relationship between the pulmonary artery and bronchus is used. Specifically, first, the center point, diameter, and wall thickness of the bronchus are estimated from the diameter of the pulmonary artery. The probabilities of bronchial lumen and bronchial wall thickness are calculated for the pixels near the estimated bronchial center point. A filter image is obtained using the probability and the pixel value of the original image. Details are described below.

  The medical image processing apparatus 100 reads the cross-cut section defined in step ST2d (step ST3a). Then, the medical image processing apparatus 100 calculates the pulmonary artery diameter from the cross-cut section (step ST3b). This diameter is defined as A.

  Subsequently, the medical image processing apparatus 100 calculates the bronchial inner diameter B by multiplying A calculated in step ST3b by a coefficient s shown in FIG. 5A (step ST3c). The coefficient s is arbitrarily set in the range of 0.65 <s <0.7.

The medical image processing apparatus 100 then subtracts the calculated bronchial inner diameter B from the estimated bronchial outer diameter (bronchial diameter) D and multiplies it by 1/2 as shown in Equation (1). The wall thickness T is calculated (step ST3d).

  At this time, in the process of step ST3d, it is necessary to calculate the outer diameter D of the bronchus. As shown in FIG. 5B, the bronchus outer diameter and bronchial wall thickness are determined in advance as T = 0.2 * D. Therefore, T = 0.2 * D and Equation (1) The bronchial wall thickness T can be calculated by solving the simultaneous equations.

The medical image processing apparatus 100 calculates the bronchial outer diameter D by multiplying the calculated bronchial wall thickness T by 2 and adding the bronchial inner diameter B, and the center point k of the pulmonary artery already obtained above, Using bronchial outer diameter D and pulmonary artery diameter A, bronchial center point Ci is calculated by equation (2) (step ST3e).

  In other words, as shown in FIG. 6, all points (distance on the dotted line in FIG. 6) that are a distance (A + D) / 2 from k are candidates for the bronchial center point (0 <= i <= 2π). ). The medical image processing apparatus 100 performs the process shown in FIG. 7 on the center point Ci of each bronchus on the circumference of FIG. 6 (on the dotted line in FIG. 6) by changing i from 0 to 2π. Specifically, in the process shown in FIG. 7, a straight line is radially extended from the central point Ci of each bronchus, and pixels existing on the straight line from the central point Ci of each bronchus are determined as filter processing targets. Then, among the determined pixels to be subjected to filter processing, pixels within a distance D / 2 from the bronchial center point Ci are determined as processing targets. Hereinafter, it is referred to as a processing target pixel.

The medical image processing apparatus 100 calculates the multivariate probability distribution function P (x) for all the processing target pixels (step ST3f). Here, a normal distribution is treated as an example. Note that other probability distribution functions may be used for the multivariate probability distribution.

  Moreover, the following etc. are mentioned as an example of a variable vector.

(A1) Pixel value of the original image in the processing target pixel
(A2) Distance value between target pixel and bronchial center point Ci
(A3) Pixel value of the original image at a distance (B + T) / 2 from the bronchial center point Ci
(A4) Pixel value of the original image at a position-(B + T) / 2 away from the bronchial center point Ci When using the above four variables, the dimension d is 4 and the average vector μ can be defined as follows: .

Furthermore, the variance-covariance matrix S can be defined as follows:

The medical image processing apparatus 100 calculates the filter value F (x) of each pixel from the existence probability obtained in step ST3f (step ST3g). The filtered pixel value of each pixel is defined by adding the pixel value of the original image multiplied by a constant probability. Note that integration may be used instead of addition.

Here, the constant α can be given, for example, as follows (refer to FIG. 8 for Tr1, Br, and Tr2). Tr1 and Tr2 indicate the bronchial wall, and Br indicates the bronchial lumen.

  For reference, FIG. 9 shows the result of performing filter processing on pixels on a certain straight line. A curve represented by a solid line on the graph is a profile of the original image. Filter 1 (dotted line in FIG. 9) and Filter 2 (dashed line in FIG. 9) are obtained by changing the value of α in the equation (7). Also, the black circle marker pixels filled in on the graph are the bronchial lumen pixels, and the white circle marker pixels are the bronchial wall pixels. Therefore, the effect of the filter can be confirmed if the value of the black circle marker is smaller than that of the original image and the value of the white circle marker is larger than that of the original image.

  FIG. 9 shows that the filter effect is reflected in both Filter1 and Filter2. Furthermore, a color map image can be created from the existence probability obtained by the expression (3). Similarly to FIG. 9, when attention is paid to pixels on a certain straight line and a color map image of the existence probability relating to the bronchial lumen is generated, FIG. 10 is obtained. In FIG. 10, the vertical axis indicates the value of the existence probability, and the horizontal axis indicates the pixel positions of the bronchial lumen and the bronchial wall. Here, 3 to 7 are the pixel positions of the bronchial lumen, and the vicinity of 1 to 3 and 7 to 9 are the pixel positions of the bronchial wall.

  As in FIG. 9, black circle markers filled in the graph indicate bronchial lumen pixels, and white circle markers indicate bronchial wall pixels. It can be seen that the existence probability value is high at the black circle marker pixel.

  The cross-cut section shown in FIG. 4 is an image in which the bronchial region is emphasized in white as shown in FIG. 11 by applying the bronchus emphasis filter.

  As described above, according to the above-described embodiment, the medical image processing apparatus 100 uses the information on the pulmonary artery running along the bronchi to improve the accuracy of bronchi extraction, and the medical image database obtained by the medical image photographing apparatus 200. From the image data stored in 300, a pulmonary artery region having a contrast difference from the surroundings is extracted from the bronchial region, the existence probability of the bronchial region is obtained from the pulmonary artery region, and the bronchial region is emphasized based on the existence probability Therefore, improvement in segmentation accuracy of the bronchial region can be expected, thereby reducing diagnostic errors between observers. In addition, since the image processing time can be reduced, the diagnostic speed is improved.

  Further, according to the above embodiment, the bronchial region is extracted from the enhanced image generated by the image generating unit 140 by the bronchial region extracting unit 160, and this bronchial region is superimposed on the image data and displayed on the display unit 150. The image can be easily observed, and the improvement of the diagnostic speed can be expected.

  Furthermore, according to the above embodiment, since the image generation unit 140 generates a color map image based on the presence probability calculated by the presence probability calculation unit 130, the observer looks at the color map image. Quantitative diagnosis of bronchial wall and bronchial stenosis can be performed with high accuracy.

(Other embodiments)
In the above embodiment, the individual processes described in the pulmonary artery extraction unit 120, the existence probability calculation unit 130, the image generation unit 140, the display unit 150, and the bronchial region extraction unit 160 can be processed by software using a computer program.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

  DESCRIPTION OF SYMBOLS 100 ... Medical image processing apparatus, 110 ... Input part, 120 ... Pulmonary artery extraction part, 130 ... Existence probability calculation part, 140 ... Image generation part, 150 ... Display part, 160 ... Bronchial region extraction part, 200 ... Medical image imaging device, 300: Medical image database.

Claims (7)

A pulmonary artery extraction unit that extracts a pulmonary artery region from a medical image;
From the pulmonary artery region extracted by the pulmonary artery extraction unit, an existence probability calculation unit for calculating the existence probability of the bronchial region,
Based on the presence probability calculated by the existence probability calculation unit, to that medical image processing apparatus and a weighted image generation section that generates an image emphasizing the bronchial area. The medical image processing apparatus further comprising Ru請 Motomeko 1 wherein a display unit for displaying an image generated by the enhanced image generating unit. Further comprising a bronchus extraction unit for extracting a bronchial region using the image generated by the enhanced image generation unit;
The display unit, the bronchial region extracted by bronchus extraction unit, the medical image processing apparatus to that請 Motomeko 2 wherein the display superimposed on the medical image. A map image generation unit that generates a color map image based on the presence probability calculated by the presence probability calculation unit;
Wherein the display unit, the medical image processing apparatus 請 Motomeko 2 wherein that displays a color map image generated by the map image generation unit. The existence probability calculation unit, the medical image processing 請 Motomeko 1, wherein you calculated on a straight line extending in a radial shape from the center point of the center point and bronchial region estimated from the diameter of the extracted pulmonary artery extraction unit area apparatus. A medical image processing method used in a medical image processing apparatus,
The medical image processing apparatus extracts a pulmonary artery region from a medical image;
The medical image processing device calculates the existence probability of the bronchial region from the extracted pulmonary artery region,
The medical image processing apparatus, based on said existence probability, medical image processing method that generates an image emphasizing the bronchial area. A program executed by a medical image processing device, wherein the medical image processing device extracts a pulmonary artery region from a medical image;
From the pulmonary artery region extracted by the pulmonary artery extraction unit, an existence probability calculation unit for calculating the existence probability of the bronchial region,
Based on the presence probability calculated by the presence probability calculation unit, an enhanced image generation unit that generates an image in which the bronchial region is enhanced;
A program to be run.