JP3637049B2 - Dynamic range compression method for radiographic images - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method for compressing a dynamic range of a radiographic image, and more particularly, to an improvement technique of a method for compressing a dynamic range of a radiographic image by processing an original image signal to obtain an image signal carrying an image having a narrower dynamic range than the original image. .
[0002]
[Prior art]
Conventionally, for example, Patent Document 1 discloses a method for compressing a density region while ensuring appropriate observation of a fine structure in an image region in a radiographic image.
[0003]
In the compression method disclosed in Patent Document 1, an unsharp mask signal (blurred mask signal) Sus is obtained by averaging the original image signal Sorg in a predetermined mask region including each pixel point corresponding to each pixel point. The processed image signal Sproc having a compressed dynamic range is obtained as Sproc = Sorg + f (Sus), where f (Sus) is a function that monotonously decreases as the non-sharp mask signal Sus increases. .
[0004]
[Patent Document 1]
Japanese Patent Laid-Open No. 3-2222577
[Problems to be solved by the invention]
However, according to the conventional dynamic range compression method, the dynamic range is compressed with a certain degree of compression regardless of the dynamic range of the subject. For example, compression processing of a low density portion in a chest image of a person who is thinner than the standard In the case of overweight, the density of the mediastinum increases and the diagnostic performance deteriorates. On the other hand, in the case of a fat person, there is a problem that the mediastinum is crushed white due to insufficient compression. There was a possibility.
[0006]
The present invention has been made in view of the above-described problems, and an object of the present invention is to avoid the occurrence of excess or deficiency in the degree of compression due to differences in subjects or differences in imaging parts in a method for compressing the dynamic range of a radiographic image. And
[0007]
[Means for Solving the Problems]
Therefore in the first aspect of the present invention, the dynamic range compression method radiographic image obtaining processed image signal Sproc carrying the narrow image dynamic range than the original image by processing the original image signal Sorg, the original of the subject and dynamic range information of the image, along with changing the physician compression degree of the dynamic range according to at least one of the information of the imaging site of a subject, provided the maximum degree of compression of the dynamic range exceeds the maximum compression The setting of the degree was prohibited .
[0008]
According to such a configuration, so that the compression degree by the difference of the object is not excessive or insufficient, and the information of the dynamic range of the original image of the subject, the compression degree of the dynamic range according to at least one of the information of the imaging region of the subject Change. Also, when changing the compression rate (compression level) based on information on the dynamic range of the subject, a maximum value of the compression rate is set in advance, and a compression rate exceeding the maximum value is set. In this case, the compression rate is limited to the maximum value, and a compression rate exceeding the maximum value cannot be set.
[0009]
According to the second aspect of the present invention, the unsharp mask signal Sus is obtained corresponding to each pixel point, and the original image signal Sorg is corrected by the correction value f1 (Sus) which is a function of the unsharp mask signal Sus. The configuration is such that a finished image signal Sproc is obtained.
[0010]
According to a third aspect of the present invention, the correction value f1 (Sus) has a characteristic that the inclination changes with an increase in the non-sharp mask signal Sus.
According to a fourth aspect of the present invention, the degree of compression of the dynamic range is increased in accordance with an increase in the dynamic range of the subject.
[0011]
According to the fifth aspect of the present invention, the degree of compression of the dynamic range is changed so that the dynamic range of the processed image signal Sproc substantially matches the reference dynamic range.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below.
FIG. 1 showing an embodiment shows a radiological image reading apparatus including an image processing apparatus to which a dynamic range compression method for radiographic images according to the present invention is applied, and shows an example of imaging a human body for medical use. Show.
[0014]
Here, the radiation generation source 1 is controlled by the radiation control device 2 to irradiate the subject (human body chest or the like) M with radiation (generally, X-rays).
The recording / reading device 3 includes a conversion panel 4 on a surface facing the radiation source 1 across the subject M. The conversion panel 4 has energy according to the radiation transmittance distribution of the subject M with respect to the radiation dose from the radiation source 1. Are accumulated in the photostimulated layer, and a latent image of the subject M is formed there.
[0015]
The conversion panel 4 is provided with a photostimulable layer on a support by vapor deposition of photostimulable phosphors or coating photostimulable phosphors, and the photostimulable layer blocks adverse effects and damages caused by the environment. In order to do so, it is shielded or covered by a protective member.
[0016]
As the photostimulable phosphor material, for example, a material as disclosed in JP-A-61-72091 or JP-A-59-75200 is used.
A light beam generation unit (gas laser, solid state laser, semiconductor laser, etc.) 5 generates a light beam whose emission intensity is controlled, and the light beam reaches the scanner 6 via various optical systems and deflects there. In addition, the optical path is deflected by the reflecting mirror 7 and guided to the conversion panel 4 as stimulated excitation scanning light.
[0017]
The condensing body 8 is positioned in the vicinity of the conversion panel 4 to which the excitation light is scanned and is positioned at the condensing end as an optical fiber, and is proportional to the latent image energy from the conversion panel 4 scanned with the light beam. Receives stimulated light emission with emission intensity.
[0018]
Reference numeral 9 denotes a filter that passes only light in the stimulated emission wavelength region from the light introduced from the light collector 8. The light that has passed through the filter 9 enters the photomultiplier 10 and corresponds to the incident light. Photoelectrically converted into a current signal.
[0019]
The output current from the photomultiplier 10 is converted into a voltage signal by the current / voltage converter 11, amplified by the amplifier 12, and then converted into a radiation image signal composed of digital data for each pixel by the A / D converter 13. The
[0020]
The digital radiation image signal (original image signal Sorg) is sequentially output to the image processing device 14 incorporating a microcomputer.
Reference numeral 15 denotes an image memory (magnetic disk device) for storing image signals.
[0021]
Reference numeral 16 denotes an interface for transmitting a radiation image signal read out from the image processing apparatus 14 directly or from the image memory 15 to the printer 17.
Reference numeral 18 denotes a read gain adjustment circuit. The read gain adjustment circuit 18 adjusts the light beam intensity of the light beam generator 5, adjusts the gain of the photomultiplier 10 by adjusting the power supply voltage of the high voltage power supply 19 for photomal, and a current / voltage converter. 11 and the gain adjustment of the amplifier 12 and the input dynamic range of the A / D converter 13 are adjusted, and the reading gain of the radiation image signal is adjusted comprehensively.
[0022]
Note that the method for obtaining the original radiation image signal Sorg to be output to the image processing device 14 is not limited to a method for photoelectrically converting the photostimulated luminescence produced by scanning the photostimulable phosphor with excitation light. Alternatively, for example, a method of reading an image of a radiation film by photoelectric conversion, a method of irradiating a fluorescent material with radiation transmitted through a subject to convert it into fluorescence, and reading the fluorescence by photoelectric conversion may be used.
[0023]
The original radiation image signal Sorg may be in a form proportional to the intensity of the detected radiation, or may be in a form proportional to the logarithm of the intensity of the detected radiation, but the latter is preferable.
[0024]
Here, the image processing device 14 has an image processing function for compressing the dynamic range of the input original image signal Sorg and obtaining a processed image signal Sproc carrying an image having a narrower dynamic range than the original image. The image processing for dynamic range compression is performed according to the following equation.
Sproc = Sorg + f1 (Sus)
In the above equation, Sus is a non-sharp mask signal obtained by averaging the original image signal Sorg in a predetermined mask area including each pixel point corresponding to each pixel point.
[0025]
Further, f1 (Sus) added to the original image signal Sorg is a correction value obtained as a function of the unsharp mask signal Sus, and monotonously decreases as the unsharp mask signal Sus increases.
[0026]
Note that “monotonically decreasing” does not necessarily mean that f1 (Sus) decreases as the non-sharp mask signal Sus increases, and partially changes f1 (Sus) even if Sus changes. It is assumed that there may be a region that does not.
[0027]
FIG. 2 is a diagram showing an example of the correction value f1 (Sus). In FIG. 2, the correction value f1 (Sus) is zero in a region larger than the Sus1 point, and the unsharp mask signal Sus from the Sus1 point. It has the characteristic of increasing at a constant rate to the positive side as the value decreases.
[0028]
This characteristic is expressed by the following formula.
f1 (Sus) = β (Sus1-Sus) (Sus <Sus1)
f1 (Sus) = 0 (Sus ≧ Sus1)
In the present embodiment, an increase in the image signal Sorg indicates an increase in density. Therefore, by adding the correction value f1 (Sus) to the original image signal Sorg, the low density side is raised and the low density side is compressed. Is called.
[0029]
According to such a configuration, since the dynamic range of the low region as the average density is compressed by the unsharp mask signal Sus, it is possible to obtain an image in which the low density side is compressed while keeping the contrast of the fine structure as it is.
[0030]
Further, compression on the low density side and compression on the high density side can be performed simultaneously. In this case, the compression degree required on the low density side and the compression degree required on the high density side Therefore, it is preferable to set different compression degrees to perform compression with the optimum dynamic range on the low and high density sides.
[0031]
FIG. 3 shows an example of the characteristic of the correction value f1 (Sus) adapted to the head front image. The characteristic of the correction value f1 (Sus) shown in FIG. become.
[0032]
f1 (Sus) = β ₁ (Sus1-Sus) (Sus ≦ Sus1)
f1 (Sus) = 0 (Sus1 <Sus ≦ Sus2)
f1 (Sus) = β ₂ (Sus2−Sus) (Sus> Sus2)
Here, the coefficients β for determining the degree of compression were β ₁ = 0.6 and β ₂ = 0.4.
[0033]
Further, Sus1 and Sus2, which determine the compression density region, are weighted average weights of the maximum signal value Smax and the minimum signal value Smin when the maximum signal value of the region of interest in the image is Smax and the minimum signal value is Smin. The calculation was performed as follows by changing k.
[0034]
Sus1 = k · Smax + (1−k) · Smin (k = 0.5)
Sus2 = k · Smax + (1−k) · Smin (k = 0.9)
As described above, in the configuration in which compression is performed based on different compression levels on the high density side and the low density side, as in the case where a certain gradient (coefficient β) is given, one density region In this case, optimal compression is performed, but there is no problem that the desired compression effect cannot be obtained in the other density region, and optimal compression can be performed simultaneously on the high density side and the low density side.
[0035]
Particularly in medical radiographic images, it is more preferable to reduce the degree of compression on the high density side compared to the low density side.
In the above-described embodiment, the correction value f1 (Sus) is given as a linear function that monotonously decreases as the non-sharp mask signal Sus increases. For example, as shown in FIG. In this case, the degree of compression (gradient) meeting the respective requirements may be set on the low density side and the high density side.
[0036]
In addition, the degree of compression is not limited to the configuration that is changed by the coefficient β, and the degree of compression may be varied depending on other parameters.
By the way, since the dynamic range of the original radiation image Sorg changes depending on the subject, if the dynamic range compression is performed at a certain compression degree, the compression degree may be excessive or insufficient, and an optimal compression process may not be performed.
[0037]
Therefore, the present embodiment is configured to change the compression degree and compression method of dynamic range compression in accordance with subject information.
FIG. 5 is a block diagram illustrating a basic configuration of an embodiment in which the degree of compression is changed according to the dynamic range (subject information) of the subject.
[0038]
Here, as described above, the original image signal Sorg acquisition means A photoelectrically converts the photostimulated luminescence produced by scanning the photostimulable phosphor with excitation light and photoelectrically converts the image of the radiation film. There are a method of reading by conversion, a method of irradiating a fluorescent material with radiation that has passed through a subject, converting it to fluorescence, and photoelectrically converting the fluorescence to read.
[0039]
The dynamic range measurement means B is a means for measuring the dynamic range of the subject. For example, the maximum signal value and the minimum signal value in the subject image area are obtained from the histogram analysis of the original image signal Sorg, and the dynamic range of the subject is obtained therefrom. Determine the range.
[0040]
Further, the maximum signal value and the minimum signal value may be obtained from the profile information of the original image.
Further, the dynamic range of the subject may be detected without using an image signal.For example, the intensity distribution of the radiation transmitted through the subject may be directly measured, or the radiation may be converted into light once. The light intensity distribution may be measured, or the thickness of the subject may be measured.
[0041]
The dynamic range setting means C is a means for setting the dynamic range of the processed image signal Sproc. For example, the dynamic range setting means C may be configured to be input / set by a keyboard operation by an operator, Are stored as default values, or an operator selects from a plurality of types of default values, or the default values are changed and set by the operator.
[0042]
The compression rate determining means D is a means for variably setting the compression rate when the dynamic range is compressed by the compression means E described later.
Here, the compression rate is defined as compression rate = (dynamic range of processed image signal Sproc) / (dynamic range of original image signal Sorg), and a decrease in compression rate indicates an increase in the degree of compression.
[0043]
The compression rate determining means D can include a conversion table as shown in FIGS. 6 to 8 for converting the dynamic range of the subject into the compression rate, and determine the compression rate with reference to the conversion table. .
[0044]
Further, a ratio between the dynamic range (reference dynamic range) of the processed image signal Sproc set by the dynamic range setting unit C and the dynamic range of the subject measured by the dynamic range measurement unit B is calculated to obtain a predetermined value. Alternatively, the compression rate may be set so as to match the dynamic range of the processed image signal Sproc being processed.
[0045]
As described above, if the compression rate is changed according to the dynamic range of the subject, the compression rate is excessive or insufficient due to the difference in the physique of the subject (human body), or the dynamic range of the processed image signal Sproc is large. Variations can be avoided and a stable processed image can be provided.
[0046]
The compression means E can obtain a processed image signal Sproc having a compressed dynamic range by correcting the original image signal Sorg with a correction value f1 (Sus) that is a function of the unsharp mask signal Sus. , Sproc = Sorg−f1 (Sus), the correction value f1 (Sus) is a function that monotonously increases as the non-sharp mask signal Sus increases, and Sproc = Sorg + f1 (Sus). In this case, the correction value f1 (Sus) is a function that monotonously decreases as the non-sharp mask signal Sus increases.
[0047]
Specifically, the correction value f1 (Sus) is given as, for example, f1 (Sus) = β (Sus−Sus1).
Here, the compression ratio changes by changing the coefficient β (0 ≦ β ≦ 1.0), and in the above formula, the compression ratio = (1−β) / 1. Therefore, from the set compression ratio By obtaining the coefficient β, processing according to the set compression rate can be performed.
[0048]
Instead of determining the compression rate by the compression rate determining means D, the coefficient β may be determined directly.
In addition to the configuration using the unsharp mask signal Sus, the compression means E may perform compression by a convolution operation.
[0049]
FIG. 9 is a block diagram illustrating a basic configuration of an embodiment in which the compression ratio is changed according to the dynamic range of the subject and the compression condition is determined based on the analysis result of the imaging region (subject information).
[0050]
The same elements as those in the block diagram of FIG. 5 are denoted by the same reference numerals, and detailed description thereof is omitted.
In FIG. 9, the imaging region analysis means F is a means for analyzing the imaging region of the subject based on the original image signal Sorg. For example, a well-known imaging region analysis method such as histogram analysis, contour extraction, or profile analysis is used. Can do.
[0051]
Further, recognition using a neural network or the like may be applied, and further, a configuration in which an imaging region is analyzed by a combination of these analysis methods may be employed.
[0052]
The compression condition determining means G is means for determining various conditions in the image processing for compressing the dynamic range in accordance with the imaging region analyzed by the imaging region analysis means F. Specifically, the following conditions are set: To decide.
[0053]
a. Compression method used by the compression means E: Selects the non-sharp mask method or the convolution operation method (convolution method) according to the imaging region. b. Correction function form: For example, when the unsharp mask method is used and the signal correction function f1 (Sus) is given as f1 (Sus) = β (Sus1-Sus), the value of the coefficient β or the reference value Sus1 Is changed according to the imaging region. Note that the degree of dynamic range compression and the application range of dynamic range compression change by changing the coefficient β and Sus1.
[0054]
c. Dynamic range of processed image signal: A plurality of predetermined values of the dynamic range after processing are stored for each imaging region, and selected from the stored values according to the analyzed imaging region. Alternatively, a plurality of tables for converting the dynamic range of the subject into a compression ratio are provided for each imaging region, and the table is selected from the plurality of conversion tables according to the analyzed imaging region.
[0055]
In the above embodiment, the imaging region is obtained by signal analysis of the original image signal Sorg. However, the configuration may be such that information on the imaging region is input by an operator's operation or the like, and the configuration corresponding to this embodiment is shown in FIG. It is shown in the block diagram.
[0056]
In FIG. 10, an imaging part input means H is a means for inputting data indicating where the imaging part is in the radiographic image to be compressed, and the operator operates the keyboard to operate the imaging part. Information is directly input, or an operator selects a preset processing menu.
[0057]
In addition, it can also be set as the structure which inputs image processing conditions etc. simultaneously according to the imaging | photography part with the information of the said imaging | photography part.
In the compression condition determining means G, the compression processing method, function form, dynamic range after processing, etc. are determined according to the information on the imaging region, as in the above embodiment.
[0058]
As described above, if the compression condition is variably set based on the information on the imaging region of the subject, it is possible to cope with the change in the required compression rate according to the imaging region, and the compression rate is excessive or insufficient due to the change in the imaging region Therefore, the optimum compression process can be performed stably.
[0059]
As described above, in the case of a configuration in which the compression rate (degree of compression) is changed based on the dynamic range information of the subject, the maximum value of the compression rate is set in advance, and the dynamic range information of the subject is set. When a compression ratio exceeding the maximum value is set based on the above, it is preferable that the compression ratio is limited to the maximum value and setting of the compression ratio exceeding the maximum value is prohibited.
[0060]
In the embodiment, the mask size or frequency characteristic of the non-sharp mask is an important parameter that determines the diagnostic property of the image.
In the dynamic range compression processing, only an ultra-low frequency component corresponding to a rough change in the structure of the subject (smooth signal difference in the lung field, mediastinum, etc.) is extracted as the unsharp mask signal Sus, and based on Sus. By setting the correction value f1 (Sus), it is possible to compress the entire concentration range while maintaining fine structural changes (bones, blood vessels, etc.).
[0061]
When the mask size is small, the unsharp mask signal Sus includes not only the ultra-low frequency component corresponding to the rough change of the subject but also the frequency component corresponding to the change of the fine structure, and the unsharp mask signal Sus. By adding the correction value based on it, the change of the fine structure is canceled and the contrast of bones and blood vessels is lowered.
[0062]
On the other hand, if the mask size is large, the edge cut of the unsharp image becomes worse at the portion where the signal value changes suddenly as described above, and undesired compression occurs near the boundary between the region where compression is desired and the region where compression is not desired. It will be given.
[0063]
If the mask size is further increased, the frequency component corresponding to the rough change of the subject is also lost (in the extreme case, the image becomes completely flat), so the correction value based on the unsharp mask signal Sus. Even if is added, the dynamic range compression effect cannot be obtained.
[0064]
As a result of the study by the inventors from the above viewpoint, the size of the mask size is preferably 10 mm to 60 mm, more preferably 15 mm to 30 mm, and most preferably 20 mm to 30 mm on a life-size image. I found out.
[0065]
If the mask size is smaller than 10 mm, the frequency component corresponding to the change in the fine structure increases abruptly. Therefore, if the correction value is set based on the unsharp mask signal Sus obtained with such a mask size, the diagnosis is remarkably performed. Performance will be degraded.
[0066]
In particular, in chest images and abdominal images, if the mask size is set to 15 mm or more, Sus does not have a frequency component corresponding to a large blood vessel such as an aorta, and if the mask size is set to 20 mm or more, Sus becomes a rib or the like. The corresponding relatively low frequency, but no frequency component that does not want to lower the contrast is not included, so that an image with high diagnostic performance can be obtained.
[0067]
Here, the mask size refers to the average value of the length of the short side and the length of the long side if it is a rectangle, the length of one side if it is a square, the diameter if it is a circle, and the average value of the long and short diameters if it is an ellipse.
In addition, when describing the frequency characteristics of the non-sharp mask instead of the mask size, it is preferable that the modulation transfer function of the non-sharp mask is 0.5 or more when 0.01 cycle / mm and 0.5 or less when 0.06 cycle / mm. Preferably, it is 0.5 or more at 0.02 cycle / mm and 0.5 or less at 0.04 cycle / mm, more preferably 0.5 or more at 0.02 cycle / mm and 0.5 or less at 0.03 cycle / mm.
[0068]
In the averaging process for obtaining the unsharp mask signal Sus as in the present invention, weighting is performed according to the absolute value of the signal difference between the central pixel and the peripheral pixel in the mask area, and / or the mask area. By performing weighting according to the positional relationship between the center pixel and the peripheral pixels, it is possible to prevent deterioration of the edge sharpness of the non-sharp image in a portion where the signal value changes rapidly, and a preferable mask size range is It can be expanded from 10 mm to 60 mm to 10 mm to 80 mm.
[0069]
Furthermore, in the present invention, it is preferable that the maximum absolute value of the correction value f1 (Sus), which is a function of the unsharp mask signal Sus, is 1/8 to 1/2 of the dynamic range of the region of interest of the subject.
[0070]
For example, when the dynamic range of the region of interest of the subject is 2 digits, the maximum value of the absolute value of the compression correction amount is preferably from 1/4 digit to 1 digit.
Further, when the correction value f1 (Sus) is expressed by a linear function of the unsharp mask signal Sus like β (Sus1-Sus), a preferable range of β that determines the degree of compression by the inclination thereof is 0.2 to 1.0. More preferably, it is 0.4 to 0.8.
[0071]
If the correction amount is too small, the dynamic range compression effect does not appear. On the other hand, if the correction amount is too large, the density relationship of each region in the original image is reversed (for example, the median portion average rather than the average density in the lung field). The density becomes higher), and the image cannot withstand the diagnosis.
[0072]
For example, such a problem occurs when the slope β of the linear function is larger than 1.
[0073]
【The invention's effect】
As described above, according to the dynamic range compression method for radiographic images according to the present invention, the compression characteristics are changed based on information such as the dynamic range of the subject and the imaging region. There is an effect that it is possible to avoid an excessive or insufficient compression degree due to the difference.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a radiation image reading processing apparatus to which the present invention is applied.
FIG. 2 is a diagram showing an example of a function form of a correction value according to a non-sharp mask signal.
FIG. 3 is a diagram illustrating a characteristic example of a correction value for performing compression on a high / low density side.
FIG. 4 is a diagram illustrating a characteristic example of a correction value for performing compression on a high / low density side.
FIG. 5 is a block diagram of an embodiment in which compression characteristics are changed according to subject information.
FIG. 6 is a diagram illustrating an example of compression rate characteristics according to a dynamic range of a subject.
FIG. 7 is a diagram illustrating an example of compression rate characteristics according to a dynamic range of a subject.
FIG. 8 is a diagram illustrating an example of compression rate characteristics according to the dynamic range of a subject.
FIG. 9 is a block diagram of an embodiment in which compression characteristics are changed according to subject information.
FIG. 10 is a block diagram of an embodiment in which compression characteristics are changed according to subject information.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Radiation generation source 3 ... Record reading apparatus
14 ... Image processing device
15 ... Image memory
16… Interface
17 ... Printer

Claims (5)

In a dynamic range compression method of a radiographic image, an original image signal Sorg representing an original image based on radiographic image information transmitted through a subject is processed to obtain a processed image signal Sproc carrying an image having a narrower dynamic range than the original image.
And dynamic range information of the original image of the subject, along with changing the physician compression degree of the dynamic range according to at least one of the information of the imaging site of the subject,
A method for compressing a dynamic range of a radiation image , wherein a maximum value of the compression degree of the dynamic range is provided, and setting of a compression degree exceeding the maximum value is prohibited . The non-sharp mask signal Sus is obtained corresponding to each pixel point, and the processed image signal Sproc is obtained by correcting the original image signal Sorg with the correction value f1 (Sus) which is a function of the unsharp mask signal Sus. The method for compressing a dynamic range of a radiographic image according to claim 1 . The radiographic image dynamic range compression method according to claim 2 , wherein the correction value f <b> 1 (Sus) has a characteristic that an inclination changes as the non-sharp mask signal Sus increases. The method of compressing a dynamic range of a radiographic image according to any one of claims 1 to 3 , wherein the degree of compression of the dynamic range is increased in accordance with an increase in the dynamic range of the subject. The radiological image according to any one of claims 1 to 3 , wherein the degree of compression of the dynamic range is changed so that the dynamic range of the processed image signal Sproc substantially matches a reference dynamic range. Dynamic range compression method.

Images