JPS6244224A - Image processing method and apparatus - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of the Invention) The present invention relates to a method of image processing a plurality of medical images taken of substantially the same part of the same subject, and an apparatus for carrying out the method, and particularly to The present invention relates to an image processing method and apparatus that can be suitably used in a radiation image information recording and reproducing system using a phosphor sheet.

(Technical Background of the Invention and Prior Art) When a certain type of phosphor is irradiated with radiation (X-rays, α-rays, β-rays, γ-rays, electron beams, ultraviolet rays, etc.), part of this radiation energy is absorbed into the phosphor. It is known that when this phosphor is irradiated with excitation light such as visible light, the phosphor exhibits stimulated luminescence depending on the accumulated energy. Exhaustive phosphor (lI14I exhaustible phosphor)

Using this stimulable phosphor, radiation image information of a subject such as a human body is partially recorded on a sheet of stimulable phosphor, and this stimulable phosphor sheet is scanned with excitation light such as a laser beam to make it shine. The stimulated luminescent light is generated, the resulting stimulated luminescent light is read photoelectrically to obtain an image signal, and this image signal is subjected to appropriate signal processing (image processing) for the purpose of improving the suitability of observation and interpretation. The applicant has already proposed a radiation image information recording and reproducing system that outputs a radiation image of the subject as a visible image to a recording material such as a photographic light-sensitive material, a CRT, etc. based on an image signal that has been subjected to signal processing. . (JP-A-55-12429, JP-A-56-11395, etc.) Such a radiation image information recording and reproducing system can be used for various medical diagnoses of the human body and the like.

For example, by using this system, approximately the same part of the same subject is photographed multiple times at predetermined intervals, and the reproduced images (outputted as visible images on the recording material or CRT, etc.) are compared with each other to create a new image. It is possible to perform a so-called comparative diagnosis to diagnose whether or not a particular lesion has occurred, or to diagnose the degree of change in the lesion, such as the extent to which the lesion has healed or worsened between the previous and current lesions. can.

However, even if a plurality of images are taken of substantially the same region of the same patient, the density and contrast of each image may differ if the images were taken on different days.

This is because it is extremely difficult to perform photography under completely the same photographing conditions or under a stimulable phosphor sheet having completely the same sensitivity and characteristics if the photographs are taken on different days.

If the density and contrast change for each image as described above, the comparative diagnosis described above, which diagnoses the occurrence and degree of change of a lesion based on changes in normal density and, in some cases, changes in density and contrast, will be hindered. This causes a problem in that proper diagnosis becomes difficult.

Note that in this specification, density means brightness when a reproduced image is displayed on a CRT.

(Object of the Invention) In view of the above-mentioned circumstances, the object of the present invention is to provide a method for image processing a plurality of medical images taken of the same part of the same subject so as to facilitate accurate comparative diagnosis. The object of the present invention is to provide a device for carrying out the method.

(Structure of the Invention) In order to achieve the above object, the image processing method according to the present invention provides an image processing method based on image information of the entire image of a plurality of medical images to be image processed. or based on information on both images of substantially the same image portions that do not include the lesion portion of each of the plurality of medical images to be image processed, that is, image portions of substantially the same imaging site. The present invention is characterized in that gradation processing is performed on each of the plurality of medical images so that at least the densities of substantially the same image portions that do not include the lesion portion among the plurality of medical images match.

Further, in order to achieve the above object, the gradation processing device according to the present invention also includes an image memory that stores image information of a plurality of medical images to be subjected to image processing, and each image memory that stores image information in the memory. A histogram creation unit that creates an image information histogram of the entire image or a substantially identical image portion that does not include the lesion portion for each medical image; and a histogram creation unit that creates an image information histogram of the entire image or a substantially identical image portion that does not include the lesion portion, and compares each histogram and generates an image information histogram of the entire image or approximately the same image portion that does not include the lesion portion between each medical image. and a gradation processing section that performs gradation processing on each medical image so that at least the densities of the same image portions match.

That is, the present invention, for example, when performing image processing on two images that are the targets of comparative diagnosis, uses the entirety of both images or a substantially identical image portion (hereinafter, this image portion is referred to as a region of interest) that does not include a lesion area. Both images are subjected to gradation processing so that at least the densities of the two images match.

Note that in the present invention, it is sufficient to perform image processing so that at least the densities match, but it is more preferable to shift the image processing so that the densities and contrasts match. Alternatively, image processing may be performed so that the histograms of image information (image signals) of both regions of interest match.

(Effects of the Invention) As described above, the method according to the present invention is capable of image processing a plurality of medical images taken of substantially the same part of the same subject, the entire image of each of the plurality of medical images or an image of a region of interest. The method is characterized in that, based on the information, gradation processing is performed on each of the plurality of images so that at least the density of the entire image or the region of interest among the plurality of images matches.

The respective regions of interest in the plurality of images are image portions of substantially the same region of the same subject, and are portions that do not include a lesion portion. Therefore, if each image is subjected to gradation processing (image processing) based on the image information of these regions of interest so that at least the densities of those regions of interest match,
The non-lesional areas in each image after processing have the same density,
Therefore, if the density of the lesion differs, the difference in density directly indicates the degree of deterioration or healing of the lesion.

Furthermore, in general, the proportion of the lesion area in the image is not that large, and therefore, in most cases, processing the image so that at least the density of the entire image matches, means that in most cases, at least the density of the region of interest in the image is the same. It can be equated with image processing to match.

Therefore, if a plurality of images are image-processed using the image processing method and apparatus according to the present invention, the degree of change in the lesion area between the plurality of images can be expressed extremely accurately on the images, and comparisons can be made extremely easily and accurately. The effect of being able to perform a diagnosis is achieved.

(Embodiments) Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

In the embodiment described below, the two images 10 and 12 shown in FIGS. Area A1A' is a substantially identical image area that does not include the lesion area (area of interest), and areas B and B' are the lesion area.

In this embodiment, when performing image processing on both of the above images,
Image processing is performed so that at least the densities of both regions of interest A and A' match.

First, the above-mentioned regions of interest ASA' are set.

Both regions of interest A and A' can be set anywhere and in any size as long as they do not include the lesion areas B and B', but of course, both regions of interest 1*A and A' can be set as shown in the figure. It needs to be approximately the same part in the image. The above-mentioned region of interest is set manually, for example, based on information regarding various imaging conditions such as the region to be imaged that can be known in advance.

Next, both interest areas iii set in this way! A, A'
Based on the image information of , both images 10 and 12 are processed so that at least the densities of the regions of interest A and A' match.

Image processing so that the densities match means that both regions of interest ta
Characteristic values of concentration of A and A', for example, maximum and minimum values of concentration,
This means performing gradation processing so that the average value, median value, etc. match. These density characteristic values may be determined in any manner, but for example, a density (image signal level) histogram may be created for each of the regions of interest A and A', and the density characteristic values may be determined based on the histogram. Note that the histograms of both regions of interest A and A' are nothing but image information of both regions of interest A and A' described above.

In the present invention, as mentioned above, it is desirable to perform image processing so that not only the density but also the rough contrast match. For example, image processing is performed so that the above-mentioned histograms of both regions of interest A and A' match. executed by This point will be explained below.

There are, for example, three patterns of differences in density (image signal level) histograms in both regions of interest as shown in FIGS. 2 to 4. Note that in FIGS. 3 and 4, the horizontal axis is the image signal level, and the vertical axis is the frequency, similarly to FIG. 2.

The difference pattern shown in FIG. 2 is the histogram h of region of interest A and region of interest iii! The histogram h' of A' is shifted in the direction of the image signal level axis, and the difference pattern shown in Fig. 3 is that both histograms h and h' are similar and shifted in the direction of the image signal level axis. The difference pattern shown in Figure 4 is that both histograms h and h
' are different without any relation.

The difference shown in FIG. 2 is mainly due to the difference in sensitivity of the radiation aperture and the stimulable phosphor sheet between the two images.

The differences between FIG. 3 and FIG. 4 are mainly due to differences in tube voltage (radiation energy distribution), materials of the top plate of the cassette and bed, and presence/absence and type of grid for removing scattered radiation. This point will be explained in more detail as follows. That is, the relationship between the energy of radiation and the radiation absorption coefficient of a substance is as shown in FIG. In other words, as the energy of radiation increases for both bones and soft tissue, the absorption coefficient tends to decrease (contrast decreases from an image perspective). However, the rate of decrease is greater in bones than in soft parts. Therefore, images taken with high-energy radiation have low contrast, that is, the width of the histogram is narrow. In addition, images taken with low-energy radiation have high contrast,
In other words, the width of the histogram becomes wider. At this time, depending on the internal structure of the subject to be photographed, the image may appear as shown in FIG. 3 or as shown in FIG. 4. That is, when the object is mainly composed of one type of substance, the image becomes as shown in Fig. M3, and when the object consists of two or more types of substances with different absorption coefficients, it becomes as shown in Fig. 4.

As mentioned above, there are two possible ways to match two different histograms h1h'; one is to match one histogram to the other, and the other is to match both histograms to a certain standard histogram. This is a method of matching.

Hereinafter, a specific example of matching both histograms using the former method (matching one to the other) will be described for the three different patterns, but the latter method can also be executed using the same concept.

First, the case shown in FIG. 2 will be explained. Now, the image signal of histogram h is 5ai (i=1...K), and the image signal of histogram h' is 3bi (i=1...K).
...K), and if the gap between both histograms is ΔS, then in order to match the histogram h to the histogram h', the above 3ai may be converted to 3a i' according to the following formula.

3a i' -=3a i+△S Note that the conversion method may be performed by calculating as in the above equation, or by creating a gradation conversion table as shown in FIG. You can convert it. FIG. 9 is a block diagram showing a method of creating a gradation conversion table in the table creation section and converting Sai to 3ai' using the table.

Next, the case shown in FIG. 3 will be explained. Now histogram h
The image signal of 5ai (i = 1...K), the maximum value and minimum value of the image signal of both histograms h and h' are as shown in the figure.
axb.

3m1nb, histogram h is histogram h
To adjust to ', Sai is changed to 3ai' according to the formula below.
You can convert it to .

Sa i' =p-3a i +qS max
b-S minb Smaxa-3m1na Note that, as in the case of Fig. 2, the above calculation may be performed for conversion, or the conversion may be performed by creating a gradation conversion table as shown in Fig. 7. .

Further, the above p and q can be derived from the following two equations.

Next, the case shown in FIG. 4 will be explained. Gradation information of both images @Shi (Si (i=1.2...K))
Let the frequency of each image signal S1 in the histogram hSh' of both images be Ha(Sr)(i=1°2,...
・・K), Hb (Si) (i=1.2...
K). At this time, if the total number of pixels in the image is N, then the following equation is obtained. In both histograms h and h' as described above, in order to make one histogram h match the other histogram h', the image signal of the histogram h must be converted using the gradation conversion table GT created by the method shown in FIG. Good.

The method for creating the above tone conversion table is as shown in FIG. Examine whether or not the inequality holds, and if it holds, set GT (St ) = Sx in step ■; if not, set j = j + 1-2 in step ■, return to the above step ■, and repeat step ■. Examine whether the inequality holds. If it holds, set GT (Sl) - 82 at step ■, and if it does not hold, proceed to step ■ → step IV until it holds.
→Repeat step ■. In this way, the inequality at step ■ holds true under i=1, and at step ■ GT (81
If >=Sj, then in step V it is considered whether or not i=K holds; in this case, since i=1, it does not hold;
Therefore, in step ■, i = i + 1 = 2, and under this i = 2, again by steps ■, ■, ■, GT (82) = S
j, and the above steps are repeated until i=K, and the creation flow ends at step VN.

Figure 8 shows the gradation conversion table G created in this way.
S9J is a diagram showing the image signal before conversion, S
indicates the image signal after conversion.

Next, one embodiment of the apparatus according to the present invention shown in FIG. 11 will be described. The apparatus shown in FIG. 2 is for carrying out the method of image processing the two images described above,
In this device, image information (image signal) of one image is stored in one image memory 20, that of the other image is stored in the other image memory 22, and both histogram creation units 2
In 4.26, based on the image signals stored in both image memories 22 and 24, the entire image of each image or a substantially identical image portion that does not include the lesion area (area of interest)
A histogram of the image signal is created. Of course, the histogram creation units 24 and 26 separately contain information such as whether to create a histogram of the image signal for the entire image or for the region of interest, and if it is a region of interest, where the region of interest is. is input. Both histograms created by both histogram creation units 24 and 26 in this manner are input to a histogram analysis means 28, in which both histograms are analyzed to determine at least the density of the entire image or the region of interest, preferably The gradation processing conditions for both images are determined so that the density and contrast match. Note that determining the gradation processing conditions is synonymous with creating the gradation conversion table described above. Once the gradation processing conditions for both images are determined in this way, the gradation processing conditions for both images are determined based on the gradation processing conditions for both images.
The image signals stored in the gradation processing means 30
．． The gradation process is performed at 32.

These histogram analysis means 28 and both tone processing means 3
0.32 constitutes the gradation processing section 34.

The gradation-processing conditions determined by the histogram analysis means 28 are intended to match the overall density of both images or at least the region of interest. If desired, the gradation processing conditions determined by the analysis means 28 are added to the gradation processing conditions for that purpose, and both gradation processing means 30.32
Gradation processing is performed in .

In the embodiment shown in FIG. 11, the output from the histogram analysis means 28 is input to both gradation processing means 30 and 32. This is configured so that the histogram analysis means 28 can carry out a method of matching both histograms with the standard histograms described above. For example, the method of matching one histogram with the other histogram described above may be adopted. In this case, the output from the histogram analysis means 28 can be configured to be input only to the gradation processing means that performs gradation processing on the image signal related to one of the histograms.

The image signals subjected to gradation processing by the gradation processing means 30.32 are input to CRTs 36.38, respectively, and
Displayed as a visible image on RT36.38.

Although the embodiments of the present invention described above all perform image processing on two images, the present invention can also be applied to a case where three or more images are subjected to image processing. Of course, each image may be subjected to image processing based on the image information of the entire image, for example, the histogram of the image signal of the entire image, instead of the image information of the regions of interest A and A'.

Furthermore, the present invention is applicable to image processing in systems other than the radiation image information recording and reproducing system using the above-mentioned stimulable phosphor sheet.

[Brief explanation of the drawing]

FIGS. 1<a) and (b) are diagrams each showing an image to be subjected to image processing according to the present invention, and FIGS. 2 to 4 are diagrams each showing an example of a difference pattern of an image signal histogram.
Figure 5 is a diagram showing the difference in the absorption coefficient of radiation energy between bone and soft tissue, and Figures 6 to 8 are gradation levels when matching the histogram h shown in Figures 2 to 4 with histogram h', respectively. 9 is a block diagram showing a method for performing gradation conversion using the gradation conversion table shown in FIG. 6; FIG. 10 is a flowchart showing a method for creating the gradation conversion table shown in FIG. 8. , FIG. 11 is a block diagram showing one embodiment of an image processing apparatus according to the present invention. 10.12... Medical image 20.22... Image memory 24.26... Histogram creation section 34... Gradation processing section Fig. 1 Fig. 11 Fig. 2 Fig. 3 Fig. 4 Fig. 6 Fig. 7 Fig. 8 Fig. 5 Fig. 9 Fig. 1 o factor

Claims (2)

[Claims] (1) In a method of image processing a plurality of medical images taken of substantially the same part of the same subject, based on the image information of the entire image of the plurality of medical images or the substantially same image part that does not include the lesion part, An image processing method comprising performing gradation processing on each of a plurality of medical images so that the densities of the entire image or at least a substantially identical image portion not including a lesion portion match each other. (2) In a device that performs image processing on multiple medical images taken of substantially the same part of the same subject, there is an image memory that stores image information for each medical image, and for each medical image whose image information is stored in the memory. , a histogram creation unit that creates a histogram of image information of the entire image or a substantially identical image portion that does not include the lesion portion; 1. A gradation processing unit that performs gradation processing on each medical image so that at least the densities of substantially the same image portions match each other.