Density resolution test method for spiral CT machine

Abstract

The present invention provides a density resolution test method for a CT machine. The method comprises the following steps: firstly determining a window width and a window position, reading an image sequence that meets the DICOM standard and is generated by scanning a test object with the spiral CT machine, generating a grayscale histogram, linearly mapping grayscale values ​​within a certain range of the image sequence to opacity, using the mapping as a grayscale window, confirming the range value of the grayscale window according to the grayscale histogram, reconstructing the image sequence to generate a three-dimensional image, and finally observing whether a homogeneous background of the three-dimensional image disappears. By using the resolution grayscale difference as a standard, the grayscale window is gradually narrowed until the homogeneous background and a target point with a too small grayscale difference disappear together, leaving only a target material with a grayscale difference that reaches the highest resolution of the human eye, thereby greatly reducing the test error caused by different subjective judgments of different observers.

Description

Technical Field

[0001] The invention relates to the field of image analysis, and in particular to a density resolution testing method for a spiral CT machine. Background Art

[0002] Spiral CT machine generally refers to X-ray computer tomography equipment, which uses X-ray beams to effectively scan a certain thickness of a specified part of the human body, and determines whether there is a disease by analyzing the absorption of the X-ray beam by each part. It is an important medical mechanical equipment for clinical diagnosis. When used, the emitted X-rays will be converted into visible light, and then converted into electrical signals and then converted into digital information. It is transmitted to the computer and calculated and processed by professional software, and finally the information in digital form is obtained.

[0003] Density resolution, also known as lower contrast resolution, is an important parameter that affects the quality of CT images. It is defined as the ability to distinguish a certain size of detail from the background when there is low contrast between the detail and the background. In other words, it is the minimum aperture size of a circular hole that can distinguish two materials with low density differences (generally, the CT values ​​differ by 3 to 5 HU), that is, the minimum distinguishable density value.

[0004] The CT value is a unit of measurement (Hounsfield Unit) for measuring the density of a certain local tissue or organ in the human body. Air is -1000HU and compact bone is +1000HU. In other words, in theory, 2000 HU can correspond to 2000 different gray levels, but the human eye cannot distinguish such fine gray differences. Therefore, when diagnosing, different window widths and window positions are often set according to different diagnostic parts, and a section of CT values ​​is selected as the inspection target, and different grayscale displays are given to this section of CT values ​​for users to observe. Even so, some tissue details with unclear density differences in the diagnostic part are still difficult to distinguish due to factors such as noise. For such situations, performance verification of the low-contrast resolution of CT equipment is very necessary.

[0005] In the GB / T 19042.1-2003 standard, Article 6.13.2 stipulates the test method. The core means are: use a test device with clear object contrast (step, wedge or disk), place the test device in the specified position and appropriate X-ray beam; test after adjusting the settings and parameters specified in the standard, and manually count the visible contrast details distinguished from the background; the observed value should be based on the average of at least two observations and compared with the technical parameters.

[0006] In YY / T 0310-2015 "General Technical Requirements for X-ray Computed Tomography Equipment", Article 6.2.5 recommends the following test method for low contrast resolution: place a low contrast resolution detection phantom in the scanning field of view, and make the axis of the detection phantom coincide with the rotation axis of the scanning frame; select a set of CT operating conditions for scanning, and then reconstruct using a reconstruction algorithm; adjust the window width and window position, and observe through a monitor, using the diameter of the smallest distinguishable hole group as the standard.

[0007] Detection model such as Figure 1 As shown in the figure, the CT image of the detection model after being scanned by the spiral CT machine is as follows Figure 2 As shown in the figure, the conventional method requires observation with the naked eye to select the 0.5% circular hole with the smallest diameter that can be seen clearly, thus obtaining a low contrast resolution. Figure 2 It can be found that there are some ambiguous areas, and the human hole selection is subjective and cannot be accurately judged.

[0008] It can be seen that the above two test methods are based on the results of manual observation of the scan. However, different people have different abilities to distinguish the scan results of low-density differences. The 6.13 note in the GB / T 19042.1-2003 standard also states that these tests are subjective and cannot give objective and independent accuracy or be compared with other systems. In order to reduce the errors caused by this subjectivity, the standard adopts the method of averaging the observations of multiple observers. In the YY / T 0310-2015 standard, no solution is proposed to circumvent or reduce the problem of human subjective errors. Summary of the invention

[0009] The purpose of the present invention is to provide a density resolution test method for a spiral CT machine, aiming to avoid subjective errors caused by manual observation of scanning results.

[0010] In order to achieve the above object, the present invention adopts the following technical solution:

[0011] A density resolution test method for a spiral CT machine, characterized by comprising the following steps:

[0012] S1, determining the window width and window position, reading the image sequence conforming to the DICOM standard generated by the spiral CT machine scanning the object to be tested, and generating a grayscale histogram;

[0013] S2, linearly mapping the grayscale value and opacity within a certain range of the image sequence, using the mapping as a grayscale window, and determining the range value of the grayscale window according to the grayscale histogram;

[0014] S3, reconstructs the image sequence to generate a three-dimensional image;

[0015] S4, observe whether the homogeneous background of the three-dimensional image disappears? If so, execute step S5, otherwise, execute step S6;

[0016] S5, the lowest value of the grayscale window range is gradually reduced by an equal amount of the resolution grayscale difference, which is the lowest grayscale that can be distinguished by the human eye, until the homogeneous background appears. After the lowest value of the grayscale window range where the homogeneous background appears is increased by one level of the resolution grayscale difference, it is used as a standard grayscale window for testing density resolution, and step S7 is executed;

[0017] S6, the lowest value of the grayscale window range is gradually increased by an equal amount of resolution grayscale difference until the homogeneous background disappears, and the lowest value of the grayscale window range where the homogeneous background disappears is used as the standard grayscale window for testing density resolution, and step S7 is executed;

[0018] S7, analyzing the three-dimensional image.

[0019] The above technical solution can be improved as follows:

[0020] As an improvement, in step S2, the grayscale value ranges from 0 to 255, and the opacity ranges from 0 to 90%.

[0021] As an improvement, in step S5, the resolution grayscale difference is 1, 2 or 3.

[0022] As an improvement, in step S1, the image sequence is a raster image file in .dcm format; in step S3, the image sequence is reconstructed to generate a three-dimensional image, specifically:

[0023] Step S301, converting the raster image file in .dcm format into a header file in .mhd format and a CT slice set in .raw format corresponding to the header file;

[0024] Step S302 , using a visualization tool library to reconstruct the CT slice set into a three-dimensional image of the object under test according to the header file.

[0025] As an improvement, in step S1, the object to be tested is a detection phantom.

[0026] The technical solution of the present invention has the following beneficial effects:

[0027] By using the resolution grayscale difference as the standard, the grayscale window is gradually narrowed until the homogeneous background and the target points with too small grayscale difference disappear together, leaving only the target material with grayscale difference reaching the highest resolution of the human eye, greatly reducing the test error caused by different subjective judgments of different observers. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solution in this embodiment, the following briefly introduces the drawings required for describing the embodiment.

[0029] Figure 1 It is a top view of the structure of the detection model;

[0030] Figure 2 To detect the CT image of the phantom after being scanned by a spiral CT machine;

[0031] Figure 3 A software interface diagram developed for the density resolution test method of the spiral CT machine according to the present invention;

[0032] Figure 4 It is a preview of the raster image and the corresponding grayscale distribution histogram;

[0033] Figure 5 It is a schematic diagram of the linear mapping between grayscale value and opacity;

[0034] Figure 6 A schematic diagram of a cross section of software for converting a raster image file into a header file and its corresponding CT slice set;

[0035] Figure 7 To reconstruct a three-dimensional image of the object under test for the first time;

[0036] Figure 8 A schematic diagram of the human eye's ability to resolve grayscale in different ranges;

[0037] Fig. 9 It is a three-dimensional image after filtering out the homogeneous background;

[0038] Fig.10 This is a schematic diagram of the grayscale window filtering result at an ideal low-contrast resolution;

[0039] Fig.11 This is the low-contrast resolution grayscale window filtering result in actual conditions. DETAILED DESCRIPTION

[0040] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0041] According to the density resolution test method of a spiral CT machine provided by the present invention, a computer program graphical interface developed using python or other computer languages ​​is shown as follows: Figure 3 As shown, the test method specifically includes the following steps.

[0042] Step S1, determining the window width and window level, reading the image sequence conforming to the DICOM (Digital Imaging and Communications in Medicine) standard generated by the spiral CT machine scanning the object to be tested, and generating a grayscale histogram.

[0043] Specifically, the object to be measured is the patient's abdomen, such as Figure 3 to Figure 4 As shown, set the window width to 2000 and the window position to 0, click "Open Image", import the raster image file named image-000001.dcm, and the raster image preview on the left and the corresponding grayscale distribution histogram on the right will pop up.

[0044] Step S2, linearly mapping the grayscale value and opacity of the image sequence within a certain range, using the mapping as a grayscale window, and confirming the range value of the grayscale window based on the grayscale histogram.

[0045] Specifically, the grayscale window is to linearly map the lowest and highest opacity parameters from 0 to 255 grayscale levels according to the opacity parameter set by the user. Figure 5 As shown, according to the minimum and maximum grayscale ranges specified by the user, the opaque voxel points within this range are displayed.

[0046] Specifically, Figure 3 to Figure 4 As shown in the figure, in the grayscale histogram, the area close to 0 on the x-axis is the black background. This is because the homogeneous background generally occupies a relatively large pixel area in the image. Therefore, we can roughly judge from a relatively high grayscale value on the y-axis that this may be the grayscale value of the background. From the x-axis to the right, there is a small peak at a point where the x value is about 20, which is used as the initial minimum range value of the grayscale window. Since there is almost no pure white in the raster image preview on the left, the initial maximum range value of the grayscale window is directly set to 255, and finally the grayscale histogram is minimized.

[0047] Step S3, reconstructing the image sequence to generate a three-dimensional image, specifically:

[0048] Step S301, specifically, Figure 6 As shown, click the "Generate mhd file" button to convert the .dcm format raster image file into a header file sample.mhd and a CT slice set sample.raw corresponding to the header file, and pop up a dialog box to save them in a specified folder;

[0049] Step S302, specifically, click the "Select mhd file" button, select the sample.mhd file in the specified folder, and reconstruct the CT slice set sample.raw into a three-dimensional image of the object under test through the open source visualization tool library (VTK, Visualization Toolkit), such as Figure 7 shown.

[0050] Step S4, observing the three-dimensional image.

[0051] Specifically, Figure 7 As shown, it can be seen that the patient's abdomen is covered by the shadow of clothing as a homogeneous background, which affects the observation of the patient's abdomen, so step S6 is continued.

[0052] Step S5, the lowest value of the grayscale window range is gradually reduced by an equal amount of resolution grayscale difference, which is the lowest grayscale that can be distinguished by the human eye, until the homogeneous background appears. The lowest value of the grayscale window range where the homogeneous background appears is increased by one level of the resolution grayscale difference, and then used as a standard grayscale window for testing density resolution.

[0053] Specifically, in order to avoid the subjective error caused by the scanning results of the human phantom, this technology quantifies the human eye's resolution ability into a grayscale difference. At present, there have been many studies on the human eye's grayscale difference resolution ability. Taking the uint8 data type commonly used in computers as an example, the grayscale range is 0 to 225, and the human eye's resolution ability in different grayscale ranges is as follows: Figure 8 As shown,

[0054] The resolution level refers to the minimum grayscale difference that the human eye can distinguish. For example, when the grayscale is between 32 and 191, the grayscale difference that the human eye can distinguish is relatively small, about 1 to 2 grayscale levels. When the grayscale range is 192 to 255, the corresponding minimum distinguishable grayscale difference is 2 to 3 grayscale levels. Therefore, if the grayscale of the patient's abdomen is displayed in the range of 32 to 255, we may as well set the minimum grayscale difference that the human eye can distinguish to 3 as a standard.

[0055] Step S6, the lowest value of the grayscale window range is gradually increased by an equal amount of resolution grayscale difference until the homogeneous background disappears, and the lowest value of the grayscale window range where the homogeneous background disappears is used as a standard grayscale window for testing density resolution.

[0056] Specifically, when the lowest value of the grayscale window range is gradually increased to 152, the clothing shadow as the homogeneous background is filtered out. Fig. 9 shown.

[0057] S6, analyzing the three-dimensional image.

[0058] Specifically, at the ideal density resolution, the filtering result of the grayscale window is as follows: Fig.10 As shown, in fact, when the object to be measured is a detection model, some low-contrast circular holes are often confused with the homogeneous background of the detection model, making it difficult for the naked eye to clearly distinguish them. When the grayscale difference between these low-contrast circular holes and the background is lower than the set value, they will be filtered out by the grayscale window together with the background.

[0059] In summary, the grayscale window is gradually narrowed according to the grayscale difference until the homogeneous background disappears, so as to obtain a standard grayscale window for testing density resolution. The standard grayscale window can filter target points that are not much different from the homogeneous background, that is, Fig.11 As indicated by the right arrow in , these target points are regarded as details that cannot be distinguished by the human eye, such as Figure 2 In the ambiguous area, the homogeneous background of the series of circular holes with a density of 0.5% in the upper left corner can be eliminated. The observer only needs to select the target points in the remaining part to make a more accurate judgment more easily, which reduces the test error caused by the different subjective judgments of different observers to a certain extent.

[0060] Those skilled in the art will appreciate that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk, or an optical disk, etc.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be within the protection scope of the present invention.

Claims

1. A density resolution test method for a spiral CT machine, characterized in that: The steps include: S1, determining the window width and window position, reading the image sequence conforming to the DICOM standard generated by the spiral CT machine scanning the object to be tested, and generating a grayscale histogram; S2, linearly mapping the grayscale value and opacity within a certain range of the image sequence, the mapping is used as a grayscale window, and the range value of the grayscale window is determined according to the grayscale histogram; the grayscale value range is 0 to 255, and the opacity range is 0 to 90%; S3, reconstructs the image sequence to generate a three-dimensional image; S4, observe whether the homogeneous background of the three-dimensional image disappears? If so, execute step S5, otherwise, execute step S6; S5, the lowest value of the grayscale window range is gradually reduced by an equal amount of the resolution grayscale difference, which is the lowest grayscale that can be distinguished by the human eye, until the homogeneous background appears. After the lowest value of the grayscale window range where the homogeneous background appears is increased by one level of the resolution grayscale difference, it is used as a standard grayscale window for testing density resolution, and step S7 is executed; S6, the lowest value of the grayscale window range is gradually increased by an equal amount of resolution grayscale difference until the homogeneous background disappears, and the lowest value of the grayscale window range where the homogeneous background disappears is used as the standard grayscale window for testing density resolution, and step S7 is executed; S7, analyzing the three-dimensional image.

2. The density resolution test method of a spiral CT machine according to claim 1, characterized in that: In step S5, the resolution grayscale difference is 1, 2 or 3.

3. The density resolution test method of a spiral CT machine according to any one of claims 1-2, characterized in that: In step S1, the image sequence is a raster image file in .dcm format; in step S3, the image sequence is reconstructed to generate a three-dimensional image, including the following steps: Step S301, converting the raster image file in .dcm format into a header file in .mhd format and a CT slice set in .raw format corresponding to the header file; Step S302 , using a visualization tool library to reconstruct the CT slice set into a three-dimensional image of the object under test according to the header file.

4. The density resolution test method of a spiral CT machine according to claim 3, characterized in that: In the step S1, the object to be detected is a detection phantom.

Images