JPH0458941A - Photographing posture discriminating device for digital radiation image - Google Patents

Abstract

PURPOSE:To make the density of an interest area specific and uniform by comparing a signal level of a level distribution running along an image signal and plural kinds of signal level thresholds, respectively, and discriminating a photographing posture in the image signal, based on the number of different kinds of its large/small variation pattern. CONSTITUTION:Digital radiation image data inputted from an image input part 1 is stored temporarily in a frame memory 2, inputted to a signal thinning-out part 11 in an image processing part 3, passes through a contrast emphasis correcting part 12, and reaches a projection generating part 13. Subsequently, at the time of detecting a pattern of a large/small variation to a threshold, its smoothing is executed by a smoothing processing part 14 so that a minute variation of this projection is not grasped as a large/small variation for showing a feature of a photographing posture. Next, in a variation pattern detecting part 15, each value of the projection and a signal threshold of plural values are compared successively, and in a photographing posture discriminating part 16, the photographing posture is discriminated by how many the same patterns as a feature pattern shown as a feature in a specific photographing posture exist.

Description

[Detailed Description of the Invention] <Industrial Application Field> The present invention relates to a photographing body position determination device for digital radiographic images, and more specifically, the present invention relates to a photographing body position discriminating device for digital radiographic images. The present invention relates to a device that automatically determines a photographing body position in order to determine processing conditions when performing signal processing.

<Prior Art> Radiographic images such as X-ray images are widely used for medical purposes, and methods of obtaining these radiographic images as electrical image signals include, for example, the following methods.

In other words, the radiation that has passed through the human body as an object is absorbed by a certain type of phosphor, and then this phosphor is excited with light or thermal energy, thereby removing the radiation energy accumulated by the phosphor due to the absorption. It emits fluorescence and detects this fluorescence with a photoelectric conversion element to electrically obtain radiographic image information.The radiographic image signal is digitized and then subjected to gradation processing and spatial frequency processing to be displayed on a CRT, etc. It is outputted and visualized (see Japanese Patent Laid-Open Publication No. 189853/1983, etc.).

By the way, in the processing stage before outputting the digital radiation image signal as a visible image as described above, it is necessary to perform gradation processing and spatial frequency processing as described above to make a visible image suitable for interpretation. However, even if the image data is of the same part of the human body, the shooting position (orientation of the human body)
When photographing with different images, the density of the region of interest in the region may change in each reproduced image.

For example, in order to diagnose the thoracic vertebrae, when a human chest is photographed from the front as shown in Figure 5, the region of interest, the thoracic vertebrae, overlaps with the mediastinum, which is difficult to penetrate, but as shown in Figure 6, If the image is taken from the side, it will overlap with the lung field, which is easily penetrated by radiation. Therefore, if such image data is processed under the same image processing conditions and then reproduced, the frontal image will have a relatively low density in the thoracic vertebrae region, and the side image will have a relatively high density.

Therefore, various methods and devices have been proposed for automatically determining in what body position the image data was photographed, and thereby performing image processing optimal for each body position.

This method and apparatus include density histograms of image data from 1 to 1.
There are methods that use the signal level distribution along a predetermined direction of the image in image data (hereinafter referred to as projection), and methods that perform cumulative operations for each (Japanese Patent Laid-Open No. 63-262128, JP-A-61-262132 1 JP-A-63-26213
4, JP-A No. 63-262139, etc.), and those for determining the degree of separation of bimodal peaks (see JP-A No. 63-262, etc.).
129, JP-A-63-262135, etc.), those that perform function approximation (JP-A-63106642, JP-A-63-262130, JP-A-63-2)
No. 62131, JP-A No. 63-262137,
(See Japanese Patent Laid-Open No. 63-262138, etc.).

Furthermore, there are methods for determining representative values and characteristic values from density histograms (see Japanese Patent Application Laid-Open No. 63-106641, etc.), and methods for calculating variance of density histograms (see Japanese Patent Application Laid-Open No. 63-262).
133, etc.), and one that calculates the sum of signal values near the center of the projection in the left and right direction (Japanese Patent Application Laid-open No. 1983-
262136, etc.) have been proposed.

<Problem to be solved by the invention> However, with the above method, the amount of calculation required to automatically determine in which body position the image was taken is low, even when using the density histogram or when using projection. The problem was that there were too many. For example, a standard pattern for front image data and a standard pattern for side image data are obtained from a large amount of image data, and the degree of similarity between these standard patterns and data input later is examined (Japanese Patent Laid-Open No. 63-262130). Publication No., JP-A-63-2
62137, etc.), it is necessary to use complex techniques such as multiple regression analysis to obtain standard patterns.
This results in a large amount of calculation.

In addition, when determining the imaging position using the density histogram, the imaging position is determined by capturing the shape difference in the density histogram, but the density histogram of the frontal chest image data and the density histogram of the chest side image data are used to determine the imaging position. Since the difference in shape between the density histogram and the density histogram is small, there are many errors in judgment.

Furthermore, when determining the shooting position using projection, if there is a lateral positional shift of the human body within the shooting range, the projection pattern based on the image data will change even if the shooting position is the same. For,
There is a problem in that it becomes impossible to ensure accuracy in determining the shooting position.

The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to provide an apparatus that can accurately determine the imaging position of a subject in a digital radiation image with a small amount of calculation.

<Means for Solving the Problems> Therefore, in the present invention, as shown in FIG. a magnitude change pattern detection means that compares the signal level of the detected level distribution with a plurality of types of signal level thresholds, and obtains a magnitude change pattern of the signal level with respect to each threshold value along the predetermined direction of the image; photographing position determination means for determining the photographing position of a subject in the digital radiation image signal based on the number of types of size change patterns with respect to each threshold value detected by the size change pattern detection means; The body position determination device can now be configured.

Here, the photographing body position determining means may be configured to determine the photographing body position of the subject by comparing the number of patterns that are the same as a pre-stored standard size change pattern with a predetermined number.

Further, it is preferable that the photographing body position discriminating means is configured to determine the photographing body position of the subject by combining a plurality of determinations based on comparisons between the number of specific size change patterns and a predetermined number.

On the other hand, it is preferable that the level distribution detecting means detects the signal level distribution using a total value or an average value of the signal levels of each pixel column substantially orthogonal to a predetermined direction of the image in which the level distribution is to be detected.

Further, the level distribution detection means may be configured to detect the distribution of signal levels by directly using the signal levels of each pixel row along the predetermined direction of the image for which the level distribution is to be detected.

Furthermore, as shown by the dotted line in FIG. 1, a signal thinning means for reducing at least one of the number of pixels of the digital radiation image and the number of signal quantization steps is provided, so that the level distribution detecting means or the signal thinning means can be used to It is preferable to detect the level distribution based on a digital radiation image signal in which at least one of the number of pixels and the number of quantization steps of the signal is reduced.

Further, as shown by the dotted line in FIG. 1, the digital radiographic image signal is provided with a last enhancement correction means for applying contrast enhancement correction to the digital radiographic image signal so that the distribution of the density histogram of the digital radiographic image is broadened, and the level distribution is It is preferable that the level distribution is detected based on the digital radiation image signal corrected by the detection means or the contrast enhancement correction means.

Further, as shown by the dotted line in FIG. 1, a level distribution smoothing means for smoothing the level distribution detected by the level distribution detecting means is provided, so that the level distribution smoothing means can be used either by the magnitude change pattern detecting means or by this level distribution smoothing means. It is preferable to determine the magnitude change pattern based on the smoothed level distribution.

<Function> According to the photographing body position discriminating device for digital radiation images having such a configuration, the state of the distribution of signal levels in a predetermined image direction, more specifically, the state of changes in the magnitude of the signal level in the predetermined direction of the image, can be detected in one image. Rather than making a distinction by cutting at a threshold, it is possible to estimate the ups and downs in the level distribution two-dimensionally by cutting at multiple types of thresholds, and it is possible to estimate the ups and downs in the level distribution two-dimensionally. The shooting position is determined based on the number of types of patterns.

Here, when determining the shooting position of the subject based on the number of different types of size change patterns, it is necessary to compare the number of patterns that are the same as a standard size change pattern stored in advance with a predetermined number. By knowing in advance the imaging position when the level distribution exhibits many of the reference magnitude change patterns, it is possible to determine whether or not the imaging position corresponds to the reference magnitude change pattern. By combining the discrimination based on the comparison between the number of specific size change patterns and a predetermined number, it becomes possible to accurately discriminate a plurality of types of imaging positions.

When detecting the level distribution of a digital image signal in a predetermined image direction, the sum or average value of the signal levels of each pixel column substantially orthogonal to the predetermined direction of the image may be used, or the signal level distribution of the pixels along the predetermined direction of the image may be used. The signal level of each column may be used as is.

In addition, after reducing at least one of the number of pixels of the digital radiation image and the number of signal quantization steps, or applying correction to emphasize the contrast so as to widen the distribution of the density histogram of the digital radiation image, the level distribution is By detecting this, it is possible to detect a pattern of changes in size based on a small amount of data, and it is also possible to easily capture the characteristics of the photographed body position.

Further, by smoothing the detected level distribution, it is possible to prevent the determination of the imaging position from being influenced by minute level changes.

<Example> The present invention will be explained in detail below.

FIG. 2 shows an embodiment of the digital radiation image imaging body position discriminating device according to the present invention.

Here, the image input unit 1 includes a digital X-ray imaging device configured to scan and read a latent image of a radiation image stored (accumulated) on a stimulable phosphor, or a digital This may be an image scanner that converts it into a signal, or a Filink system that already records multiple pieces of digital radiation image data.

The original digital radiation image data f (x, y) input from the image input section 1 is stored in the frame memory 2 at -1 and then passed to the image processing section 3 . The image data f (x, y) (x=L 2°*+,
Nx,y=1.2.・+, Ny) is X.

It indicates the signal level of each pixel whose address is specified by y, and in this embodiment, a pixel column extending in the X direction (horizontal direction in FIG. 3) is referred to as a "row", and ) will be called a ``column''.

The image processing unit 3 is equipped with the configuration necessary for the photographing body position discriminating device according to the present invention, and also performs gradation conversion and frequency emphasis 2 to reproduce a visible image suitable for image interpretation based on the result of determining the photographing body position. It includes an output image processing section 4 that performs various image processing and analysis such as contour extraction.

As will be described in detail later, the characteristic pattern storage unit 5 stores a reference pattern which is data necessary for determining the photographing body position, a threshold Th for determining the number of such patterns, a threshold T for determining the signal level, etc. I remember.

An image output unit 6 receives the image data processed by the output image processing unit 4, and records the image data (digital radiation image signal) after image processing in the filing system again or reproduces it on a CRT. or recorded on film by a printer. Here, when filing the image data in the filing system after determining the shooting position of the subject (in this example, the human body),
If you record image information such as patient name, imaged part name, imaged date, etc., and record the imaged position in correspondence with the image data,
This is convenient for later searches and eliminates the need to repeatedly determine the imaging position.

Next, an embodiment of the digital radiation image imaging body position determination device according to the present invention will be described along with each of the processing units 11 to 16 included in the image processing unit 3.

The original digital radiation image data f (x, y) that has been stored in the frame memory 2 is first input to a signal thinning section (signal thinning means) 11 in the image processing section 3 .

This signal thinning unit 11 thins out the number of pixels and the number of gradations (the number of quantization steps) of the original digital image data f (x, y) to reduce the amount of image data used for imaging body position determination, and performs the following calculation. Try to reduce the amount.

Specifically, the original digital image data f(x,
y) into a plurality of small regions and find the average value or representative value of the signal value for each small region, thereby reducing the number of pixels and the number of gradations. For example, 2048 x 2464 pixels.

Original digital image data f(x
, y), even if the signal is thinned out to 128×154 pixels and 256 gradations, the imaging body position determination according to the present invention can be performed without any problem.

Note that the image data whose number of pixels and number of gradations has been thinned out by the signal thinning unit 11 is expressed as f"(x', y') (X'
=1.2. ..., Nx', y'= 1.2. ...
, Ny') (Nx'< Nx, Ny'< N
y).

However, the above-mentioned thinning process is for further reducing the amount of calculation in the imaging body position discriminating apparatus according to the present invention, which requires only a small amount of calculation, and the thinning process may be omitted.

In the next contrast emphasis correction section (contrast emphasis correction means) 12, the image data f'(x",
Contrast emphasis correction of y') is performed as necessary to obtain contrast emphasis data f''(x').

y') is obtained. This is because the signal values (shade values) of the original image data often do not have much difference in shading. This process is performed to prevent the difference in height between the peaks and valleys of a curve from becoming small, and by applying contrast enhancement correction, the density difference is emphasized and the shading of the image data related to determining the shooting position is improved. Make sure the characteristics of the change are clearly expressed.

Contrast enhancement is performed using, for example, the following known technique (see "Introduction to Computer Image Processing", supervised by Hideyuki Tamura, published by Souken Shuppan Co., Ltd.).

The level range of the signal before emphasis is [a□in+amax
l, If the signal range after emphasis is [b, , , , b, , ,], then the signal value S o r before emphasis: □ is the following ■
Contrast enhancement signal value S pr by formula or formula 0.

. Convert to

(logs oriz loga n1lo) +
b + nin'''■Here, when the contrast is emphasized using the formula (2), the density is converted linearly, and when the formula 0 is used, the density is converted nonlinearly.

In addition, as shown in the following formula (■), a density histogram with an average density value m and a standard deviation σ is expressed as an average value mn. 7. Standard deviation σ. . ,.

The signal value S o t + t before emphasis is normalized to the density histogram of the contrast-enhanced signal value S p
r. . You may also convert it to .

σ Further, by flattening the density hiss 1 to crumb so that the number of pixels in the entire density range is equal, the contrast next projection creation unit (level distribution detection means) 1
3, image data f''(x') that has been subjected to signal thinning processing and contrast 1 to emphasis correction processing.

y′) based on the projection p(x′)(image
level distribution along the direction).

The projection p (x') is the image data f
(x', y') can be used as the distribution expressed by the following equation (however, as shown in Figure 3, 1≦n1≦n2≦Ny', and the signal sampling range is n1≦y). '≦n2).

P + (x') -'): f'' (
x ' , y ' )...■ In other words, formula (■) calculates the sum of signal values for each pixel column extending in the vertical direction (X direction) of image data f''(x' , y'), and formula 0 Then, the average value of the signal values for each pixel column extending in the vertical direction of the image data f''(x.y') is determined. In some cases, it is also possible to use the sum or average value of signal values for each pixel row extending in the left-right direction (X direction) of the image data f''(x',y').

In this embodiment, as described above, a projection is created based on the sum or average value of the signal values of each pixel column, but the signal value of each pixel in a specific pixel row is used as is to create a projection in the It is also possible to create a projection that shows signal level changes along the left and right directions. However, it is necessary to appropriately select the specific pixel row so that the characteristics of the photographing position are clearly expressed.

As shown in FIG. 4, the projection p (x') obtained by the projection unit 13 includes minute fluctuations in signal values, so a pattern of changes in magnitude with respect to a threshold T, which will be described later, is detected. Sometimes, the process of smoothing the projection p (x') is carried out by the next smoothing processing unit (level distribution smoothing means) so that this minute change is not interpreted as a change in size that indicates the characteristics of the shooting position. ) 14.

The smoothing process is performed, for example, by the projection p(X')
This can be done by replacing the point of interest within the area with the average value within its neighborhood. As shown in Figure 4, the position of each point in the projection p (x') in the horizontal direction is x
′, the signal value is a function p (x′), so
When performing neighborhood processing for n left and right neighbors, the value p' (i) after smoothing at position i can be expressed by the formula p' (i) - Σ p (x')/ (2xn+1). By performing this process multiple times, small fluctuations in signal values as shown in FIG. 4 can be smoothed out. Note that the smoothing process in the smoothing processing section 14 can also use the median value within the neighborhood.

Next, in the change pattern detection section (size change pattern detection means) 15, the smoothed projection p
By sequentially comparing each part of '(x') with multiple types of signal thresholds below, a pattern of changes in the signal level for each threshold T is obtained, and in the next imaging position discriminating section (imaging body position discriminating means) 16, The photographing body position is determined based on the number of patterns that are the same as the characteristic pattern that characteristically appears in a particular photographing body position in this variation pattern.

That is, in the case of the front and side views of a human chest radiographic image, the fifth
In the distribution of signal values in the left-right direction of the images shown in straight lines in Figures and Figure 6, in the frontal image inside the shell, there is a valley in the signal level representing the thoracic vertebrae, which is difficult for radiation to penetrate, in the center in the left-right direction. , there are peaks of signal level on both sides indicating the lung field through which radiation can easily pass, and furthermore, there are valleys of signal level at the left and right ends representing the thoracic region through which radiation is difficult to pass. On the other hand, in the side image, there are peaks at the left and right edges of the signal level indicating the clear area of the radiation, and near the center there is a valley indicating the structure.
Characteristic waveforms are hard to appear on the sides of the body. These characteristics are substantially the same even if the distribution of the sum or average value of the signal values of each pixel column in the direction perpendicular to the straight line is taken.

Here, when the signal level of the projection and the threshold value T are sequentially compared while changing the signal level threshold value T,
For each part of the projection, a variation pattern in which a part smaller than the threshold value T continues, then changes to a larger part, and continues to be a smaller part again is sampled by the number of threshold values T, and these variation patterns are shown in Figures 5 and 6. As is obvious, the shooting position varies greatly depending on whether the photograph is taken from the front or from the side.

In this way, the patterns that appear more often in frontal images and the patterns that appear more often in side images can be identified, so if you find the characteristic patterns that correspond to these shooting positions in advance, you can easily adjust the input image data. The photographing body position can be determined based on the number of patterns that are the same as the characteristic pattern among the patterns corresponding to the threshold value T obtained by comparing the projection and the threshold value T.

Detection of such a size change pattern and determination of the photographing body position will be specifically described below.

Change pattern detection section (size change pattern detection means) 15
Then, each point of the projection p'(x') is compared with the threshold T in order from the end, and for example, if the value on the projection is larger than the threshold T, 1 is given, and if it is smaller, 0 is given. Here, the length of 1 or 0, ie, ■ and O
Ignoring the run length of , assigning the code 1 to the run 1 and the code 0 to the run 0, the repeating pattern of codes l and O represents the fluctuation of the waveform of the projection with respect to the threshold T. There will be. Furthermore, by changing the threshold T in preset steps, the above-mentioned 1.0 sign can be set over the entire range of signal values.
The state of the waveform in the projection can be captured two-dimensionally, and many variation patterns of 1 and 0, which are characteristic of the projection, will appear.

For example, as shown in FIG. 7, when the threshold value T is changed at intervals of ΔT in the signal value range t0 to t6, as shown in Table 1 below, tn≦'I '＜t,
+1 (n=0.t, ..., 5) in each section,
The feature patterns with the code 0 and the code l and the number of feature patterns are obtained.

Table 1 When such processing is performed on each of the radiation image data shown in FIGS. 8 to 12, (01010) is shown in FIG. 8, which is a frontal image, and (101) is shown in FIG.
), In Fig. 10, which is a front right-shifted image, (10101
), and (10) in Figure 11, which is a lateral right-shifted image, and (1010) in Figure 12, which is a frontal image of a child or a thin person.
101), and only a few patterns other than the above are present in each image. Therefore, (01010).

The patterns (10101) and (1010101) are the characteristic patterns for the front image (standard size change pattern for the front), and the patterns (101) and (10) are the characteristic patterns for the side image (standard size change pattern for the side). If this is stored in the feature pattern storage unit 5 in advance, information on which patterns the input image data projection has most of, and the features corresponding to the shooting positions stored in the feature pattern storage unit 5 can be stored in the feature pattern storage unit 5 in advance. By comparing the information with the pattern information, the photographing position discriminating unit (photographing position discriminating means) 16 determines whether the image is a photographed image of a human body from the front or a side view. It is something that can be determined.

It is preferable that the signal level threshold T used for determining the projection waveform described above is determined in advance for each characteristic pattern and stored in the characteristic pattern storage section 5 together with the characteristic pattern. When making the determination, as shown in the flowchart of FIG.
5), and the determination may be made based on whether the human image data has a certain feature pattern or more than a predetermined number, or it may be possible to perform the discrimination based on whether the human image data has a certain number of feature patterns or more. The photographing body position may be determined by comparing them together with threshold values (Th6, 7).

In the determination shown in the flowchart of Fig. 14, standard front (
(see Figure 8), standard side (see Figure 9), front right deviation (see Figure 10), side right deviation (see Figure 11), front small (see Figure 11),
(See FIG. 12) By individually determining the number of each feature pattern, the five types of imaging positions can be determined.

That is, the flowchart in FIG. 14 includes pre-processing for determining the shooting position, in which original image data is input at Sl, signal thinning processing is performed at S2, and processing for further enhancing contrast is performed at S3. I do. Then, in step S4, a projection is created based on the image data subjected to these processes, and in the next step 85, a process of smoothing this projection is performed.

In S6, the smoothed projection signal level and the threshold value T are compared while changing the threshold value T, and a pattern of changes in the signal level is detected for each threshold value T.

The number of patterns of change in size for each of the detected threshold values T is determined to determine the number of patterns that are the same, and is used for determining the photographing body position from S7 onwards.

In S7, the number of patterns that are the same as (01010), which is a pattern that characteristically appears when the body is in the standard frontal posture as shown in FIG. 8, is compared with the threshold value of the number of patterns, Thl, and Th1 When the pattern (01010) appears above, it is determined in S8 that the photographing position is the standard frontal position as shown in FIG.

Similar determinations are made in steps 89 to S1.6, in S9 the number of standard side feature patterns (101) is determined, in S11 it is determined whether the number of front right deviation characteristic patterns (10101), and in S13 It is determined whether the number is the right-shifted characteristic pattern (10), and further, in S15, it is determined whether the number is the front small characteristic pattern (+, 01.0101),
When the number of images exceeds the respective thresholds Th2 to Th5, it is determined that the image is taken in the body position of the corresponding feature pattern (any of the normal side, front right shift, side right shift, and small front position) (S 10.12. 14.16).

Then, the number of patterns that are the same as the patterns characteristically appearing in the above five types of photographing positions is set to a threshold Thl~5.
When smaller than the front standard.

It is determined in S17 that the photographing position is not one of the standard side view, front right deviation, right side deviation, or small front position, and is "other".

On the other hand, in the determination shown in the flowchart of FIG. 15, the determination is simplified and the imaging body position is determined into two types, front and side, regardless of the presence or absence of positional deviation. That is, pre-processing 321-
For 326, refer to 8 in the flowchart in Figure 14.
1 to S6, but S27 in determining the shooting position
, the number of patterns corresponding to the standard front, the number of patterns corresponding to the front right shift, and the number of patterns corresponding to the small front are added, and the sum of the patterns that appear when all of these are front, or the predetermined When the threshold value Th6 is exceeded, it is determined in S28 that the photographing position is the front without distinguishing between standard front, front right deviation, and small front.

Then, when it is determined in S27 that the number of frontal patterns is less than the threshold Th6, in S29 the sum of the number of patterns corresponding to the standard side surface and the number of patterns corresponding to the rightward deviation of the side surface is compared with the threshold Th7, and When exceeding Th7, S30 is used without distinguishing between standard side and side deviation to the right.
It is determined that the shooting position is on the side.

If it is determined in S29 that the subject is not in the shooting position or on the side, S3
1, it is determined that the photographing position is "other", neither front nor side.

Furthermore, as shown in FIG. 13, even if the image data is a special case in which one lung has been removed, the characteristic pattern (010) is added to the set of characteristic patterns in the characteristic pattern memory 5. In addition, by comparing the number of characteristic patterns corresponding to such a human body with one lung with a threshold value of the number of patterns, it is also possible to determine the body position in which the human body with one lung is photographed from the front.

When the photographing position is determined as described above, finally, a signal representing the photographing position determination result is sent to the photographing position discriminator 16.
The image is then output to the output image processing section 4.

Then, the output image processing unit 4 directly takes in the original image data f (x, y) in the frame memory 2, performs processing such as gradation conversion and frequency emphasis using a conversion method that suits the shooting position, and then creates an image. Data F (x, y) is obtained and this image data F (xy) is output to the image output section 6.

The image output unit 6 records the image data F(xy) processed according to the shooting position from the output image processing unit 4 in a filing system, reproduces it on a CRT,
It is recorded on X-ray film by a printer. Here, since the image output unit 6 processes the image data in accordance with the imaging position of the image, the reproduced image can be made into a visible image suitable for interpretation.

<Effects of the Invention> As explained above, according to the digital radiation image imaging body position determination device according to the present invention, a wide variety of image data can be processed without being affected by the size or positional deviation of the subject in digital radiation image data. The photographing position of the subject can be easily and accurately determined from the photographic position, and based on the determination result of the photographing position, the optimal image processing for the photographing position can be performed to uniformize the density of the region of interest in the reproduced image. This has the effect of improving medical diagnosis based on radiographic images.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing the configuration of the present invention, Fig. 2 is a system block diagram showing an embodiment of the invention, Fig. 3 is a diagram showing how to represent image data in the above embodiment, and Fig. 4 is a line diagram showing the state of projection before smoothing processing in the same example as above; FIGS. Diagram 8 showing how a size change pattern is detected based on projection in the embodiment
Figures 1 to 13 are diagrams for explaining the differences in signal level change patterns depending on the imaging position, and Figures 14 and 115 are flow charts showing the content of imaging position discrimination based on the differences in the size change patterns. -It is a chart. l...Image input section 2...Frame memory 3...
・Image processing section 4...Output image processing section 5...Characteristic pattern storage section 6...Image output section 11...Signal thinning section 12...Contrast I/emphasis correction section
13... Projection creation section 14... Smoothing processing section 15... Change pattern detection section 16
・Photographing position determination unit

Claims (8)

[Claims] (1) A level distribution detection means for detecting a level distribution along an image predetermined direction of a digital radiation image signal, and a signal level of the level distribution detected by the level distribution detection means and a plurality of types of signal level thresholds, respectively. a magnitude change pattern detection means for comparing and determining a magnitude change pattern of the signal level for each threshold value along a predetermined direction of the image; and a magnitude change pattern for each threshold value detected by the magnitude change pattern detection means according to type. 1. An imaging body position discriminating device for digital radiographic images, comprising: an imaging body position discriminating means for discriminating the imaging body position of a subject in a digital radiographic image signal based on the number of images. (2) The photographing body position determining means determines the photographing body position of the subject by comparing the number of patterns that are the same as a pre-stored reference size change pattern with a predetermined number. 1. The digital radiation image imaging body position determination device according to 1. (3) The photographing body position determining means determines the photographing body position of the subject by combining a plurality of determinations based on comparisons between the number of specific size change patterns and a predetermined number. Digital radiographic imaging body position determination device. (4) The level distribution detecting means detects the signal level distribution using the sum or average value of the signal levels of each pixel column substantially perpendicular to the predetermined direction of the image for which the level distribution is to be detected. 4. A digital radiation image imaging body position determination device according to claim 1, 2, or 3. (5) The level distribution detecting means detects the signal level distribution by directly using the signal level of each pixel row along the predetermined direction of the image whose level distribution is to be detected. 3. The digital radiation image imaging body position determination device according to any one of 3. (6) Signal thinning means for reducing at least one of the number of pixels and the number of signal quantization steps of the digital radiation image, and the level distribution detection means reduces the number of pixels and the number of signal quantization steps by the signal thinning means. 6. The digital radiation image imaging body position determination apparatus according to claim 1, wherein the level distribution is detected based on a digital radiation image signal in which at least one of the following is reduced. (7) comprising a contrast enhancement correction means for applying correction to the digital radiation image signal to emphasize contrast so that the distribution of the density histogram of the digital radiation image is widened;
Claim 1, 2 or 3, wherein the level distribution detection means detects the level distribution based on the digital radiation image signal corrected by the contrast enhancement correction means.
7. The digital radiation image imaging body position determination device according to any one of 4, 5, and 6. (8) Level distribution smoothing means for smoothing the level distribution detected by the level distribution detecting means, wherein the magnitude change pattern detecting means detects the level distribution smoothed by the level distribution smoothing means. Claims 1, 2, and 3 characterized in that a change pattern of magnitude is determined.
, 4, 5, 6, or 7.