KR100333678B1 - Method for obtaining subtraction by rib segmentation in chest images - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

The present invention relates to a method of obtaining a difference image from which ribs are separated from a chest image.

2. The technical problem to be solved by the invention

The present invention provides a method of obtaining a difference-free image having a low error through image conversion after separating ribs from a chest image taken at a time interval, and a computer-readable recording medium recording a program for realizing the method. For that purpose.

3. Summary of Solution to Invention

The present invention, the first step of separating and removing the rib from the chest image taken at the time interval; And a second step of obtaining a difference image by image-converting the ribbed chest image.

4. Important uses of the invention

The present invention is used in the field of medical image processing.

Description

Method for obtaining subtraction by rib segmentation in chest images}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to medical image processing technology, and more particularly, to a method for obtaining a difference-free image in a chest image and a computer-readable recording medium having recorded thereon a program for realizing the method.

Image processing has been applied in many different fields since it was applied to transmission photography in the 1920s. In particular, medical image processing began with application to X-ray X-ray imaging, and various medical imaging devices such as computerized tomography, magnetic resonance tomography, ultrasound diagnostics, photon emission computed tomography, and positron emission tomography in the 1970s. Its importance is increasing day by day. Such medical imaging apparatuses enable to synthesize 3D stereoscopic images of organs inside the human body by providing a tomography image of the human body, and also a lot of researches have been conducted on the virtual surgery simulation to reduce the risk of surgery in relation to virtual reality.

Recently, the interest in automatic diagnosis through image information processing as well as visualization of 3D image synthesis, virtual surgery simulation, etc. has been greatly increased, and the field of computer aided diagnosis has been actively studied. In particular, research is being conducted in various applications such as automatic detection of lung tumors for early diagnosis of lung cancer, extraction of mutations over time of images, measurement of lung volume for pulmonary function test, and body fat measurement for obesity test.

Computer-aided diagnostics using time-dependent variance extraction is a method that utilizes differential images of chest X-rays taken at timed intervals for computer-aided diagnosis. Clear changes in the process can be useful for early diagnosis and development of the disease. That is, when it is necessary to read a large amount of regular X-ray (X) X-ray images, such as regular check-ups, by showing the abnormal part clearly, it can help to make an accurate diagnosis in a short time. In particular, it is a method that can help in early diagnosis of lung cancer by showing a difference image according to the time course of the X-ray image regularly taken in regular checkups for diseases that are difficult to diagnose early, such as lung cancer.

However, anatomy such as lungs and ribs according to various conditions in X-ray imaging, that is, imaging conditions such as X-ray intensity and imaging time, printing conditions, patient's posture, breathing state and heart rate, etc. Because the location, size and shape of the enemy structures are greatly different, the difference image by simple image difference does not help diagnosis.

Therefore, in order to assist in diagnosis, an image processing technique is necessary to match the overall brightness and contrast between two images and to obtain a difference image by matching the position and size of anatomical structures such as ribs and collarbones. In addition, the difference in size and location of the lungs should be obtained according to the difference image. Various image conversion methods are used for such correction.

As an example of the prior art, the contents of a patent registered as A.Kano and K.Doi in the United States on October 25, 1994, Patent 5,359,513 (Method and system for detection of interval change in temporally sequential chest images) As follows.

The conventional patent relates to a method for assisting diagnosis by detecting a change over time by obtaining a difference image from a chest X-ray image taken at intervals of time. In this patent, the following procedures are used to obtain anatomical structures such as lungs, ribs and collarbones, which appear on chest X-ray images.

1) Chest X-ray images of patients taken at timed intervals are obtained in digital format.

2) In this image, the anatomical structures such as lungs, ribs, and collarbones should be positioned to coincide with each other.

3) Choose a template of the appropriate size in one image to resolve discrepancies in the anatomical structures between the two images.

4) Measure the position where the correlation function is maximized by moving the template in the search area of another image and calculating the correlation function.

5) Repeat the template 4 times by repeating the template halfway from the selected image.

6) Select the template for the whole area to get the degree of inconsistency of the anatomical structures between the two images.

7) The measured discrepancies are approximated by a two-dimensional polynomial function.

8) Apply the nonlinear transformation to the image using the 2D polynomial obtained.

9) Find the difference of the abnormal part over time by obtaining the difference image between the deformed image and the undeformed image.

As described above, in the conventional image conversion method, a method of simultaneously converting a lung and a rib is used. This conventional method has a problem that the size and position of the ribs and the lungs do not coincide with each other because the lungs and the ribs are differently deformed by the patient's posture and breathing as different anatomical structures.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problem, and after separating ribs from a chest image taken at intervals of time, a method for acquiring a low difference image through image conversion and a program for realizing the method are recorded. Its purpose is to provide a computer readable recording medium.

1 is an exemplary view of a hardware system to which the present invention is applied.

2 is a flowchart illustrating an embodiment of a method for obtaining a difference image from which ribs are separated from a chest image according to the present invention;

Figures 3a and 3b is an example of a chest x-ray image taken at time intervals.

3c and 3d are examples of images showing lung and rib boundaries.

Figures 4a and 4b is an example of the image of the ribs removed from the image shown in Figures 3a and 3b.

4C is an exemplary view of an image obtained by image conversion in order to fit the image of FIG. 4A to the image of FIG. 4B.

4D is an exemplary view of a difference image obtained from FIG. 4C.

* Explanation of symbols for the main parts of the drawings

1: microprocessor 2: main storage

3: auxiliary storage device 4: input / output device

In order to achieve the above object, the present invention provides a method for acquiring a difference-less image from a chest image taken at a time interval, the method comprising: separating ribs from a chest image taken at a time interval to remove the ribs; step; And a second step of obtaining a difference image by image converting the ribbed chest image.

In addition, the present invention, the first function for separating and removing the rib from the chest image taken at the time interval in a computer, in order to obtain a difference image less error in the chest image taken at a time interval; And a computer-readable recording medium having recorded thereon a program for realizing a second function of acquiring a difference image by image-converting the ribbed chest image.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First, looking at the gist of the present invention through an example as follows.

In the present invention, the ribs and lungs are separated from the chest X (x) images taken at time intervals to obtain an image of the lungs from which the ribs are removed, and the ribs and the lungs are removed from the lung image. Suggest ways to reduce errors that result from inconsistencies in change. The present invention also helps to find the abnormal region covered by the ribs by obtaining an image of the lungs from which the ribs are removed from the chest x-ray image.

To this end, in the present invention, when the difference image is obtained using a chest X-ray image taken at time intervals, the lungs and ribs are separated to make the brightness of the ribs darker by the difference between the average values of the brightness of the lungs and the ribs. After removing the image of the rib, the difference image is obtained by using an existing image conversion method.

1 is an exemplary view of a hardware system to which the present invention is applied.

First of all, the system is basically used to perform algorithms that separate the lungs and ribs from chest X-rays taken at time intervals and then acquire the difference-free images through image transformation. Briefly described with reference to 1 is as follows.

As shown in FIG. 1, the hardware system to which the present invention is applied includes an input / output device 4 for inputting / outputting data with an external user and a lung in chest X-ray images taken at intervals. Main / secondary storage device (2,3) and the main / secondary storage device (2,3) for storing necessary data in the process of acquiring the difference-free image through image conversion after separating the frame and the ribs. Microprocessor (1) for controlling the / output device (4) and for obtaining the error-free difference image through image conversion after separating lung and rib from chest X-ray image taken at time intervals It is provided.

Through such a system, after separating the lung and the rib from the chest X-ray image taken at time intervals, a method of obtaining a difference-free image through image conversion is performed. When the program including the process of 2 is embedded in the microprocessor 1 and the chest X-ray image is input and the program is executed, the chest X-ray photographed at a time interval by the program is executed. After separating the lungs and the ribs from the image, the difference-free image with less error is obtained through image transformation. This detailed operation will be described with reference to FIGS. 2 to 4.

2 is a flowchart illustrating a method of obtaining a difference image in which ribs are separated from a chest image according to the present invention. 3A and 3B are examples of chest X-ray images taken at time intervals, and FIGS. 3C and 3D are examples of images showing lung and rib boundaries, and FIGS. 4A and 4B are views of FIGS. An example of an image with ribs removed from the images shown in FIGS. 3A and 3B, FIG. 4C is an example of an image converted to fit the image of FIG. 4A to the image of FIG. 4B, and FIG. 4D is a difference obtained from FIG. 4C. As an example of an image, a circle represents a lung tumor.

In order to obtain a difference image as shown in FIG. 2, first, a chest X-ray image taken at time intervals is obtained (21). 3A and 3B show an example of a chest X-ray image taken at time intervals. The size of the image is 210 pixels in width and 450 pixels in length, and each pixel has a 12-bit value. The image of FIG. 3A is an image of a patient taken on March 27, 1998, and the image of FIG. 3B is an image of the same patient taken on March 18, 1997, and the two images are about one year apart. Has

Thereafter, the boundary between the lungs and the ribs is extracted from the chest X-ray images obtained at the time intervals (22). At this time, the boundary between the lung and the ribs is obtained by using various known boundary detection methods such as a Sobel operator method or a threshold method. 3C and 3D show the lung and rib boundaries obtained by this method.

Thereafter, images of the lungs and the ribs are separated using the obtained borders of the lungs and the ribs (23). Then, several pieces of images of lungs and ribs are obtained. Here, the image of the rib is an image in which the rib and the lung overlap. Removing these effects from the ribs alone can produce images of lungs. Looking at the process in detail as follows.

The brightness in a simple X-ray image is proportional to the density. In other words, dense parts such as bones appear bright, while low-density parts such as air appear dark. The same is true for images of the lungs and ribs. That is, the lung image appears relatively darker than the rib image. If the density of the ribs is constant, the images of the lungs and ribs overlap will be brighter with a certain brightness than the lungs. In consideration of this, if the average value of the brightness of the separated rib image and the average value of the brightness of the lung image are obtained and the brightness of the rib image is reduced by the difference, an image of the lung of which the rib image is removed can be obtained (24). If this is expressed by a formula, it is as following (Equation 1).

P _R → P _R- (m _R -m _L )

Where P _R is the pixel value in the ribs, and m _R and m _L are the average values of the pixel values in the ribs and lungs, respectively. In this way, the rib image is removed from the chest X-ray image (24). Chest X-ray images with the ribs removed are shown in FIGS. 4A and 4B.

As can be seen in Figure 4 the image of the lungs obtained by the above-described method, the size, location and shape of the lungs are different from each other due to the photographing conditions, the patient's posture, breathing and heart rate. In order to correct such a difference, various image conversion processes are performed.

First, in order to match the contrast, the brightness of one image is changed as shown in Equation 2 so that the average and standard deviation of the two images are the same.

P _A → (P _A -m _A ) x (σ _B / σ _A ) + m _B

Here, P _A , m _A and sigma _A represent the brightness, mean and standard deviation of the image A to be converted, and m _B and σ _B represent the mean and standard deviation of the brightness of another image B. Through this conventional conversion process, the contrast between the two images is the same.

Then, in order to correct the difference between the position and the shape of the anatomical structure of the two images with the same contrast, the image is grabbed from the image A to be converted into a template of a suitable size, and the search region is twice the size of the template in another image B. Grab and find the location that most closely resembles the template.

Find the position value for the entire area of A while moving the image A to be converted while overlapping the templates in half. By fitting the position data obtained in this way into a continuous function and transforming the image using the continuous function obtained by fitting, it is possible to reduce the inconsistency of the position and shape of the anatomical structure.

FIG. 4C shows an image converted by using a function obtained by finding a template having a size of 32 pixels and a height of 32 pixels in a search area having a width of 64 pixels and a length of 64 pixels and fitting a 2-dimensional fifth order polynomial function. It was. By minimizing the inconsistency of anatomical structures in this way, and obtaining the difference images of the two images (26), the image changes over time can be clearly seen.

4d shows the difference image obtained by this method. The small circles in FIG. 4D represent lung tumors. There is no lung tumor in FIG. 3b, but there is a lung tumor in FIG. 3a, which is shown in the difference image of the two images. 3A and 3B alone, it is not easy to find a lung tumor, but it can be easily found in FIG. 4D showing a difference image. As such, using the difference image, it is possible to clearly see the occurrence and progress of the disease over time.

Above has shown a specific embodiment of the present invention, the numerical values or methods used herein may vary according to a simple X (x) ray image. Importantly, in the present invention, when a difference image is obtained from a chest X-ray image, a good difference image cannot be obtained because the lungs and ribs are obtained by image conversion up to now, but the ribs and lungs are separated in the present invention. Remove the rib image and convert the image using only the lung image to obtain a good difference image.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes can be made in the art without departing from the technical spirit of the present invention. It will be apparent to those of ordinary knowledge.

The present invention as described above, when the difference image is obtained by using a simple x-ray (x) X-ray image of the chest has an effect that can significantly reduce the error caused by the mismatch of anatomical structures such as ribs and lungs.

In particular, the conventional method to obtain the difference image by image conversion of the ribs and lungs at the same time, the present invention has the effect of reducing the effect of the ribs to remove the ribs and to obtain a difference image by image conversion of the lung only.

In addition, the abnormal part located at the border portion of the ribs and lungs does not appear well before removing the ribs, but in the present invention, the abnormal part located at the boundary parts of the ribs and lungs is also well displayed by removing the image of the ribs, thereby reducing errors in diagnosis. There is an effect that can help.

Claims (6)

In the method for obtaining a difference-free image from the chest image taken at intervals of time, A first step of separating and removing ribs from the chest image taken at intervals of time; And A second step of image-converting the ribbed chest image to obtain a difference image Differential image acquisition method ribs are separated from the chest image comprising a. The method of claim 1, The first step is, Obtaining a chest image photographed at the time intervals; Extracting a boundary between the lung and the rib from the chest image obtained at intervals of time; A fifth step of separating the images of the lungs and the ribs using the obtained borders of the lungs and the ribs; And A sixth step of obtaining a lung image by removing an image of the rib from the separated chest image Differential image acquisition method ribs are separated from the chest image comprising a. The method according to claim 1 or 2, The process of removing the rib image, A method of obtaining a difference image in which ribs are separated from a chest image, characterized in that the brightness of the ribs is darkened by the difference between the average values of the brightness of the lungs and the ribs to remove the images of the ribs. In order to obtain a difference-free image from a chest image taken at intervals, A first function of separating and removing ribs from a chest image taken at intervals of time; And A second function of image-converting the ribbed chest image to obtain a difference image A computer-readable recording medium having recorded thereon a program for realizing this. The method of claim 4, wherein The first function, A third function of acquiring a chest image photographed at the time intervals; A fourth function of extracting a border between a lung and a rib from the chest image obtained at the time intervals; A fifth function of separating the images of the lungs and the ribs using the obtained borders of the lungs and the ribs; And A sixth function of obtaining a lung image by removing an image of the rib from the separated chest image A computer-readable recording medium having recorded thereon a program for realizing this. The method according to claim 4 or 5, The process of removing the rib image, The function of obtaining a lung image by removing the image of the rib by darkening the brightness of the rib by the difference between the average values of the brightness of the lung and the rib. A computer-readable recording medium having recorded thereon a program for realizing this.