JPS60246188A - Method for processing radiant ray picture - Google Patents

Abstract

PURPOSE:To obtain a picture excellent even in details by multiplying a prescribed range of a convolute function at each reproduced picture element, and applying the convolute arithmetic adding the value as to a prescribed picture and changing the spatial frequency response of the said function in response to the strength of an original picture signal to correct the picture. CONSTITUTION:A prescribed convolution function hx in an x direction as shown in Fig. (b) is matched to a picture elements of, e.g., I-th and J-th of an original picture 61, convolution functions between the picture element 62 and the function hx, hx1, hx2, hx3..., e.g., 60 functions, are multiplied and each value is added so as to attain the convolution operation. This is applied similarly as to the picture elements 63, 64... and when the picture element 61 is finished, then a prescribed convolution function hy in (y) direction in Fig. (c) is used, the convolution operation is applied similarly so as to attain picture elements 62', 63', 64' in Fig. (b) newly. This processing is applied to the entire picture so as to obtain a processing picture. Thus, this system is different largely from the modulation transfer function at a high spatial frequency from the two-dimension convolution, but the visually equal performance is attained. Further, the number of times of operations is decreased.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a radiation image processing method mainly used for medical diagnosis.

(Prior Art) In recent years, radiation images are often converted into electrical signals and processed and reproduced.

The radiation image referred to here refers to an image formed by irradiating a subject with high-energy electromagnetic waves such as radiation, gamma rays, and neutron rays, and detecting the attenuation of the electromagnetic waves by the subject two-dimensionally. For example, there are radiographs using radiation, images obtained by two-dimensionally scanning the amount of light emitted from a fluorescent phosphor and converting it into an electrical signal, or images obtained using a so-called indirect photography television set.

The processing performed in these processes corrects image deterioration caused by radiation detection elements, radiation tubes, imaging methods, etc., and/or reproduces images that match human visual characteristics to provide images with better diagnostic performance. That is.

Broadly speaking, these processes can be divided into two categories: correcting deterioration in frequency response caused by the system used to obtain the image, such as the radiation tube in the case of film, movement of the subject, phosphor for sensitization, film, etc.; A method of obtaining visually easy-to-see images by emphasizing in the spatial frequency domain according to the visual characteristics of There are two main types of methods: determining a range and increasing the contrast in that density region. That is, there are two types: spatial frequency processing and gradation conversion processing.

Conventionally, in order to perform emphasis in the frequency domain, the formula (
Method 1) is used.

D=Dorg+β(Dorg-Dus)...(1) Here, β is the emphasis coefficient, Dorg is the original image signal, and Dus
is an unsharp mask signal of the original image signal Dorg.

In this method, in order to emphasize spatial frequencies above a certain level, a non-sharp mask signal Dus corresponding to the low spatial frequency components of the original image is obtained at each scanning point, and the high spatial frequency components are obtained by subtracting it from the original image signal Dorg. The original image signal DorH
This is achieved by adding .

When an original image filter is formed using this method, Fig. 1(a), (
A filter like b) can be realized. That is,
Range a, which is the spatial frequency domain.

It is a filter that changes greatly in b and not much in other areas, and the rate of change in ranges a and b is only that of the characteristics of the unsharp mask signal Dos.

Therefore, in order to realize an arbitrary type of filter for correction of the image acquisition system using this method and correct the average transfer function of the image acquisition system, it is necessary to use a plurality of unsharp mask signals Dus as shown in equation (2). Although it is necessary to calculate this, the calculation time becomes extremely long and is not practical.

D=Dorg+β1 (Dorg-Dusl) 10β
2 (Dorg-Dus2) 4e I+ War...◆(
2) Alternatively, to achieve this without using the unsharp mask signal Dus, it is better to use equation (3). That is, as shown in FIG. For the first pixel 2 in the vertical direction, I
, the result is the product of all the convolution functions for all images centered on J, and the result is the sum of the multiplications.

A new image 2 is created by performing a so-called two-dimensional convolution operation.
′ is obtained. This method is performed for all pixels of the original image l.

As shown in FIG. 2, this method requires a very large number of calculations.

As an example, if the number of vertical pixels of the image is M, the number of horizontal pixels of the image is N, the vertical size of the convolution function is K, and the horizontal size of the convolution function is L, then KXL multiplications per pixel. , addition operations and access to image information are required, which is (KXL) x (MXN) times for the entire image.

Example work lf, X-ray image cr [combination, K=lO~60. L=l
If O~60, M=lOOOO~2000, then in the maximum case, 60X60X2000X2000=144X
100xlODa.

The frequency characteristics of the system for obtaining radiation images usually take the form shown in Figure 3, and in order to correct this frequency characteristic, a fourth
If the spatial filter has frequency characteristics that change smoothly as shown in the figure, when the average transfer function of the system is corrected, the characteristics will be as shown in Figure 5, which is optimal.

That is, the important frequency components O to l of the X-ray image.

The response of the system is approximately 1.0 in the range of O p/m m. On the other hand, with a simple filter as shown in FIG. 1, correction cannot be performed with such precision.

By the way, I found that this filter expressed the fine details of the image well, but the graininess was noticeable. and,
The reason for this was found to be that in radiographic images, most of the image information is at a spatial frequency of 0.511 p/mm or less, and at spatial frequencies higher than that, although there is image information, the proportion of noise is large.

(Purpose of the Invention) This invention was made in view of the above circumstances, and it solves the above-mentioned problems. The purpose is to provide an image processing method.

(Structure of the Invention) In order to achieve the above object, the present invention scans a radiation image, reads out the radiation image information, converts it into an electrical signal, and then reproduces it as a visible image. A convolution operation in which each pixel is multiplied by a convolution function in a predetermined range and the respective values are added is performed on a predetermined image, and the frequency response of the convolution function is changed in the spatial frequency domain to correct the image. It is characterized by:

(Example) Hereinafter, an example in which the present invention uses X-rays as radiation will be described in detail with reference to the accompanying drawings.

FIG. 6 shows a method of convolution correction.

The purpose of this method is to reduce the number of convolution operations ΣΣh(text, m)XD(I+text, J+m) in X-ray images, and to realize practically any filter.

As a spatial filter in the X-ray image, Dx (I,
J) =Σhx (1)XD (I +4.

J) Dxy (I, J) = Σhy (cross) XDx (I
.

The performance equivalent to the two-dimensional convolution formula (3) described above is achieved by convolution calculation in the horizontal and vertical x and y directions (Jjubun).

In this method, for example, the first pixel 62 in the horizontal direction and the third pixel 62 in the vertical direction of the original image 61, as shown in FIG. L
The process is not performed for all of /2 and -72 to /2 in the vertical direction, but only in the horizontal X direction and the vertical Y direction with one pixel 62 as the center.

That is, by combining a predetermined convolution function hx in the horizontal X direction shown in FIG. 6(b) on the pixel 62, the convolution functions hxl, hx2 .
For example, multiply hx3 by 60 and add the respective values to perform a convolution operation.

Then, pixels 63, 64, . . . in FIG. 6(a)
When this is completed for the original image 61, each convolution function hyt is then calculated using a predetermined vertical convolution function hy in FIG.
, hyz, h3F3--, pixels 62, 63, 64
Convolution operation is performed in the same way for
I'm getting a fight. This process is performed on the entire image to obtain a processed image.

According to this method, although the modulation transfer function differs greatly in the high spatial frequency range from the two-dimensional convolution equation (3), the performance is visually equivalent to that of the two-dimensional convolution equation (3).

Also, according to this method, the number of calculations is (LtK)X (
For example, when calculated in the same way as above, L
Assuming that = = 60, it will be approximately 1730 or 48 seconds.

Here, a specific method of setting the filters hx and hy will be explained.

First of all, it is necessary to know the frequency response of the single image acquisition system,
In the radiography system, it is approximately represented by one of the curves a to d in Fig. 4, so be sure to select one as appropriate.
Alternatively, even if the characteristics of the image processing system are changed, almost the same result can be obtained.

f: lp/mm 2a: Frequency at which the response becomes 1/e e: Low natural logarithm, and the reciprocal of the frequency response value is the response of the filter.

Here, the filter shown in FIG. 7 is obtained, but with an appropriate Q
In the part where the filter response is higher than (1, 5 to 6.0), the Q value is set as B, or the spatial frequency is gradually increased to 3 to 51 p/mm as shown in C value. It may be set to 0. There are various combinations of hx and by, but here we will explain the case where hx=hy, that is, the frequency components of the image are approximately equal in X and Y.

Next, the square root of the frequency response obtained in FIG. 7 is taken, and this is taken as the frequency response of hx and hy, which is shown in FIG.

Furthermore, this is subjected to inverse Fourier transform, and the counts of hx and hy in the real domain are calculated.

The above calculations may be performed each time curves a, b, c, and d in Figure 3 are selected, but preferably they are calculated in advance and stored in ROM or floppy disk. It is better to store the information and simply read it out, which saves time for calculations.

Furthermore, it is known that the value of Q varies from person to person depending on the preference of the person making the diagnosis, but a value of about 2.0 to 4.0 is good.

FIGS. 9(a) and 9(b) are graphs of the judgments of 20 doctors with various changes in Q.

O is easier to judge than the original (+1) Δ is the same as the original (0) × is harder to judge than the original (-1) The score is expressed as follows.

The frequency response of the filter thus obtained is shown in FIG.

I found that this filter expressed the fine details of the image well, but the graininess was noticeable.

The reason for this was found to be that in X-ray images, most of the information per image is at a spatial frequency of 0.51 p/mm or less, and at spatial frequencies higher than that, although there is image information, the proportion of noise is large. .

The present invention is characterized in that the frequency response of the convolution function is changed in the spatial frequency domain to correct the image.

That is, as shown in FIG. 10, 0.5 to 1.01 p/m
At spatial frequencies of m or more, good results can be obtained by weakening the degree of emphasis and emphasizing in the spatial frequency domain.

In the above explanation, Q was set to obtain the frequency fP in FIG.
p/mm, and the frequency response at that time is Q,
It has been found that it is better to gradually lower the frequency like B or C at higher frequencies.

However, depending on the image obtaining system, correction may not be possible unless the maximum value of Q of this spatial frequency response is tripled or more.

In this case, always use this kind of filter.

Unnatural artifacts may occur in images (e.g. signal size and density of stomach X-ray images using barium contrast agent)
If the above processing is performed with the Q of the spatial frequency characteristic constant regardless of the However, the roughness at low image signal levels becomes noticeable and degrades the image quality.

In order to prevent the occurrence of these false images, it is necessary to
The processes shown in (b) and (c) are effective.

That is, in FIG. 11(a), in order to prevent the occurrence of pseudo images,
The Q of the spatial frequency characteristic was set according to the size of the image 11@S, and was made smaller in the weak part of the original image signal S below SL, increased in the range of signals that originally had many signals, and decreased in the strong part 52 and below. This is an example.

Curve A shows the case where it changes smoothly, and curve B shows the case where it bends at 31.32.
Almost no difference was seen between the two. In the case of this example, Q is made small to reduce image roughness in parts where the original image signal is weak. In addition, a portion 52 where the original image signal is strong
This lowers the Q and prevents the occurrence of black edges and uneven edges.

Specifically, low-signal areas and high-signal areas occupy a large part of the entire image, and these areas are not important for diagnosis, and the raw signal areas are particularly important for diagnosis, such as cholecystography and hepatography. In this case, the example shown in FIG. 11(a) is a case where it is desirable to emphasize only the raw signal portions other than these, since emphasizing noise etc. will hinder diagnosis.

Furthermore, in the case of frontal chest photography, if the characteristics of the convolution function are fixed, noise at low brightness in the spine and heart region will increase, which will be visually very noticeable and degrade the image quality, deteriorating diagnostic performance.

In this case, the noise can be suppressed by reducing the Q of the frequency response of the convolution function in the low brightness areas of the spine and heart and increasing it in the high brightness areas of the lungs.

This is the case in FIG. 11 1b).

The method shown in Figure 11 1b') is suitable for cases where diagnosis of low-intensity areas such as angiography and lymphangiography is important, and the low-intensity area does not occupy a large portion of the entire image. ing.

As an example of these, a case will be described in which the frequency response Q shown in FIG. 11(a) is used in frontal chest imaging.

The histogram of an image taken from the front of the chest has a shape as shown in Figure 12. U indicates the spine, ■ indicates the heart, and W indicates the lung field.In general imaging, we are mainly interested in the lung field W, and the spine U.

The part V of the heart is not so important, but it is said that it is enough if the shape can be deciphered to some extent. Therefore, the lower value S1 and upper value S2 of the original image signal as shown in FIG. 11(a) can be determined from this histogram. The lower value 51 or less and the upper value 32 or more do not need much correction, that is, the lower value (1!151 or less has a bad SlN, and the upper value 52 or more has a good SlN,
If it is emphasized, there is a risk that the above-mentioned dark edges and black edges may occur.

This lower value S1 is min+ (max-min)al
, al is about (0,1-0,5) KL, and the upper value S2 is m
i n+ (max-min)a2, α2 is (0,8
~1.5), and the point where the Louis product of the histogram is 30 to 50% is Sl, and the signal strength is 0.8 to 0.8 of the signal strength where the Louis product of the histogram is 80 to 90% of the whole image.
1.5 times may be set to 52.

For other parts, Sl and S2 can be determined in the same way from the histogram of that part, but it is difficult to determine which part has been imaged depending on the part, so you can use push buttons etc. , this can be achieved by allowing it to be selected.

In this way, when the value of Q was changed depending on the intensity of the image signal, the same evaluation as in FIG. 9(a) above was performed, and the results shown in FIG. 13(a) were obtained.

This example is chest xi! The example shown in FIG. 11(a) was evaluated using images. In this case Q is expressed using the maximum frequency response (Qm) of the filter.

According to this, as shown in Fig. 13(a) and (b), Q
It can be seen that the range is about 3 to 6.

In the explanation so far, the spatial frequency afp at which the filter has the maximum response has been fixed, but changing the spatial frequency fp with respect to the image signal can also be used to improve the image quality.

For example, as shown in Fig. 14, the spatial frequency fp is lowered in parts where the image signal is weak, that is, the low spatial frequency region is corrected for signals with poor SIN, and in parts with a good S/N ratio, the spatial frequency fp is lowered. This is a method that corrects up to high spatial frequencies.

Further, by changing the frequency fc in FIG. 10, that is, the frequency at which the frequency response of the filter becomes Ool, according to the image signal, it is effective to suppress the roughness of the image.

That is, as shown in FIGS. 15(a) and 15(b), the spatial frequency fC is made small at the signal level of the original image signal and the low noise ratio.
By setting the frequency fc high at medium to high signal levels with sufficient SlN, noise in the low signal level portion can be suppressed.

In addition, as shown in FIG. 12, when the spatial frequency fc is increased again in the high signal level portion, the occurrence of pseudo images is suppressed for images that rapidly change from a low signal level to a high signal level.
In addition, it is possible to sufficiently correct portions that affect diagnosis.

Furthermore, it goes without saying that even more effects can be obtained if the Q of the frequency response is changed at the same time as shown in FIG.

In the embodiment described above, it goes without saying that the order of calculations in the horizontal and vertical directions may be reversed.

Further, the correction by the convolution method of the present invention is similarly applied to the method shown in FIG.

This method uses Dxy (I, J)=Σhx (Jj)XD CI as a spatial filter in an X-ray image.
+1, J) +Σhy (k) There is.

In this method, for example, the first pixel 62 in the horizontal direction and the fifth pixel 62 in the vertical direction of the original image 61 shown in FIG. 6(a),
The convolution range is not performed for all of the nearby horizontal X direction -L/2 to L/2 and vertical direction -/2 to /2, but only for the horizontal X direction and vertical y direction centered on pixel 62. I try to do it.

That is, by combining the horizontal convolution function hx shown in FIG. 6(b) on the pixel 62, the pixel 62 and the convolution function hx
to get the result of horizontal convolution.

Similarly, the horizontal convolution function hy shown in FIG. 6(e)
Obtain the results of vertical convolution for , and add these two to form a new pixel.

This process is performed on the entire image to obtain a processed image.

Alternatively, as shown in FIG. 17, the convolution results in the horizontal direction X and the vertical direction y may be stored in the original image 61 and added later to obtain the processed image 61'.

Furthermore, the present invention is similarly applicable to the convolution method shown in FIG.

(Effects of the Invention) As described above, the present invention performs a convolution operation on a predetermined image by multiplying each pixel of each scan by a convolution function in a predetermined range and adding these respective values, Furthermore, since the image is corrected by changing the frequency response of the convolution function in the spatial frequency domain, it is possible to obtain a good image that expresses even the details of the image well and is free from roughness and has excellent diagnostic properties. I can do it.

[Brief explanation of drawings]

Figure 1 is a diagram showing an example of the characteristics of a conventional spatial frequency filter.
Figure 2 is an explanatory diagram of convolution calculation, and Figure 3 is a typical frequency characteristic diagram of a system for obtaining radiographic images. Figure s4 is a diagram showing an example of the frequency characteristics of the filter of the present invention, Figure 5 is a diagram showing the frequency characteristics when corrected, Figure 6 is an explanatory diagram of another convolution method, and Figures 7 and 8. 9 is a diagram illustrating an example of the frequency characteristic of the filter of this invention, FIG. 9 is a diagram illustrating a doctor's evaluation of the processing, FIG. 10 is a diagram illustrating an example of the frequency characteristic of the filter of this invention, and FIG. 11 is an image M412 is a diagram showing an example of a histogram of a chest X-ray image, FIG. 13 is a diagram showing a doctor's evaluation of other processing, and FIG. 14 is an example of a change in fp with respect to an image signal. FIG. 15 is a diagram showing an example of change in fc for an image signal, FIG. 16 is a diagram showing an example of filter characteristics for an image signal, and FIG. 17 is a diagram for explaining another convolution method. 1.1', 61.61'...images 2.2', 62, 63, 64.62', 63'. 64'... Pixel A... Convolution function in a predetermined range in the horizontal direction B... Convolution function in a predetermined range in the vertical direction Applicant
Konishi Roku Photo Industry Co., Ltd. Fig. 3 'Jffl/ll3L cane 1p/mm Fig. 4;,: 5 factors 10 20 All around #l$1p/mm Fig. 16 FjA Fig. 17 Y-direction real voice 9 Showa 59 June 50, 2015 Kazuo Wakasugi, Commissioner of the Japan Patent Office 1, Indication of Ik Patent application filed May 218, 1982 (1) 2, Name of invention Radiographic image processing method 3, Relationship with the amendment person case Patent applicant address: 1-26-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo Name (1)
27) Roku Konishi Photo Co., Ltd. 4 Agent Address: 151 Address No. 1, 2-23 Yoyogi, Shibuya-ku, Tokyo 6゜ Subject of amendment Scope of claims in the specification and detailed description of the invention (1)
) Amend the claims in the specification as shown in the attached sheet. (2) Change “in the spatial frequency domain” to “in the spatial frequency domain” on page 8, line 2 of the specification.
"Depending on the original image signal strength." (3) "Low luminance" on page 15, line 12 of the same book is corrected to "low signal area." (4) In the same book, page 15, line 16, "low luminance area" is corrected to "low signal area." (5) "Low brightness area" in the first line of page 16 of the same book is corrected to "low signal area." (6) Delete "low brightness" in the second line of page 16 of the same book. (7) In the seventh line of page 21 of piIJ on the right, "in the spatial frequency domain" is corrected to "according to the original image signal strength." (8) Figure 17 of the same document is corrected as shown in the attached drawing. 2. Scope of Claims (1) When scanning a radiation image, reading the radiation image information, converting it into an electrical signal, and then reproducing it as a visible image, a predetermined A radiation image processing method characterized by performing a convolution operation on a predetermined image by multiplying each convolution function of a range and adding the respective values, and correcting the image of the convolution function. (2) A patented image processing method characterized in that the maximum value of the frequency response of the convolution function is changed according to the magnitude of the image signal. (3) The spatial frequency at which the frequency response of the convolution function is maximum is determined by the image signal. The radiation image processing method according to claim 1, which is characterized in that the radiation image processing method is changed to ≧2B Procedural amendment 1 Indication of the case Patent Application No. 102387 of 1982 Title of the invention Radiographic image processing method 3 Person making the amendment Relationship to the case Patent applicant address 26-2 Nishi-Shinjuku 1-chome, Shinjuku-ku, Tokyo Name (1)
27) Roku Konishi Photo Industry Co., Ltd. 4 Agent: 151 Address: 2-23-1 Yoyogi, Shibuya-ku, Tokyo Heat explanation column 1
Drawing 7 □ty:>n@ff! IMconjJ,・ku]
h11 (1) Correct the claims in the specification as per the Pj paper. (2) In the same book, page 2, line 12, [Choxic phosphor J is corrected to ``storage phosphor.'' (3) In the third line of page 3 of the same book, "phosphor" is corrected to "fluorescent material". (4) "Expression" in lines 2 and 11 of page 7 of the same book is corrected to "amendment." (5) "Frequency response" in the first line of page 8 of the same book is corrected to "spatial frequency response." (6) In the second and third lines of page 13 of the same book, "frequency response in the spatial frequency domain" is corrected to "spatial frequency response according to the original image signal intensity." (7) In the same book, page 14, line 15, "less than" is corrected to "more than". (8) Correct rQ, 14 on page 18, line 13 of the same book to "approximately 0°1." (8) In the same book, page 21, line 7, "frequency response" is corrected to "interdirectional frequency response." (10) In the same book, page 21, line 9, ``Expression j is corrected to ``amendment''. (11) Figure 1 (a), (ri), Figure 3, Figure 4, Figure 5
Figure 10, 1 ie (a), (b). (c) and Figure 16 are replaced with the attached drawings. Above 2, claims (1) After scanning the radiation image, reading out the radiation image information, and converting it into an electrical signal, it is possible to do so! J. When reproducing as an image, a convolution operation is performed on the predetermined image by multiplying each pixel of each scan by a convolution function in a predetermined range and adding these respective values,
And the lazy frequency response of the convolution function is the original image property 1
A radiation image processing method characterized by correcting an image by changing it according to intensity. (2) A patent line image processing method characterized in that the maximum value of the 9U frequency response of the convolution function is changed according to the magnitude of the image signal. (3) The radiation image processing method according to claim 1, characterized in that the spatial frequency at which the free frequency response of the convolution function has a maximum value is changed with respect to the image signal. that's all

Claims (3)

[Claims] (1) After scanning a radiation image, reading out the radiation image information and converting it into an electrical signal, when reproducing it as a visible image, each pixel of each scan is multiplied by a convolution function within a predetermined range. A radiation image processing method characterized in that a convolution operation for adding these respective values is performed on a predetermined image, and the image is corrected by changing the frequency response of the convolution function in a spatial frequency domain. (2) A patent method characterized in that the frequency response of the convolution function is changed according to the magnitude of the image signal. (3) The radiation image processing method according to claim 1, wherein the frequency response of the convolution function changes the spatial frequency with respect to the image signal.