JP2007037864A - Medical image processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a medical image processing apparatus for compressing a dynamic range of an input image capable of obtaining high-quality and stable images in a noticed area at any time. <P>SOLUTION: The medical image processing apparatus compresses a dynamic range of an input image acquired by photography and has each following component; a region of interest-setting means (ROI analysis circuit 203) for setting the region of interest to the input image based on information of photographed site corresponding to the photography procedure, a characteristic amount extracting means (characteristic amount extracting circuit 204) for extracting the characteristic amount of pixel value composing the input image from the setup interested area, a target output value setting means (target output value setting circuit 205) for setting the target output value corresponding to the characteristic amount for each photography procedure, and a conversion process means (DRC circuit 206) for converting the pixel distribution of the input image into the pixel value for approximating the characteristic amount to the target output value. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a medical image processing apparatus. In particular, the present invention relates to an image processing apparatus suitable for dynamic range compression (DRC) processing of medical X-ray images.

  In an X-ray image, it may be difficult to express as a high-quality image over the entire area of the captured image. For example, an X-ray chest image is composed of an image of a lung field that easily transmits X-rays and an image of the mediastinum that is extremely difficult to transmit X-rays, so the range in which pixel values exist is extremely wide. . If such an X-ray image with a wide pixel value band is subjected to gradation conversion using gamma conversion, the contrast of the target of the high-frequency component contained in the image changes and the density of blood vessels and the like changes. It was an obstacle. Therefore, it was difficult to observe both the lung field and the mediastinum simultaneously with high quality images.

  Therefore, as a method for avoiding this problem, there is a dynamic range compression method. Dynamic range compression is a process that compresses the low-frequency component of the input image and expands the high-frequency component to attenuate the low-frequency component of the image and improve the contrast of the bronchi, vertebral body, trabecular bone, etc. is there.

  Regarding the dynamic range compression processing, the technique described in Patent Document 1 is known. In this technique, a pixel distribution B obtained by smoothing the pixel distribution O by a blur mask method or the like is obtained from the pixel distribution O of the original image, and k (B) obtained by density conversion of B with a predetermined parameter is obtained from the pixel distribution O of the original image. The density range is moved by subtracting. This process is expressed as equation (1) where P is the pixel distribution after image processing.

P = O−k (B) (1)
If equation (1) is modified, the following equation (2) is obtained.
P = h (B) + V (2)

Here, h (B) is a function representing the conversion of the overall gradation characteristics of the image information, as shown in the following equation (3).
h (B) = B−k (B) (3)

  Since the gradation characteristic represented by the equation (3) does not decrease in the middle of increasing monotonically, h (B) can be said to be a monotonically increasing function of B. In some cases, black-and-white reversal and dynamic range compression processing are performed together. In this case, the entire arithmetic expression is sign-reversed.

On the other hand, V in Expression (2) is expressed by Expression (4), and corresponds to a local fluctuation component of the pixel distribution O of the original image.
V = O−B (4)

JP 2003-126057 A

  However, in the above-described technique, the dynamic range is compressed with a fixed parameter determined in advance for the input image regardless of the imaging region, so that the pixel value can be converted into a desired pixel value according to the region of interest. Not considered.

  In addition, when compressing the dynamic range, a feature amount may be extracted by histogram analysis of pixel values of the entire image, and pixel value conversion output may be performed based on the feature amount. In that case, since a pixel value of a part of the histogram is extracted as a feature amount, the feature amount may be a pixel value in a different region depending on the subject. If the feature amount extraction region is different for each subject, the density of the processed image varies even if pixel value conversion processing is performed based on the feature amount. In particular, the feature amount may be extracted from a portion that is completely unrelated to interpretation, such as when the subject is a cast. In that case, the area necessary for interpretation may not be expressed with a high-quality image.

  The present invention has been made in view of the above circumstances, and one of its purposes is to always obtain a high-quality and stable image for a focused area in an image processing apparatus that compresses the dynamic range of an input image. It is an object of the present invention to provide a medical image processing apparatus that can be used.

  The apparatus of the present invention is a medical image processing apparatus that compresses the dynamic range of an input image obtained by imaging. And it comprises the following each structure, It is characterized by the above-mentioned.

Region-of-interest setting means for setting a region of interest for the input image based on imaging region information corresponding to the imaging technique.
Feature amount extraction means for extracting feature amounts of pixel values constituting the input image from the set region of interest.
Target output value setting means for setting a target output value corresponding to the feature amount for each photographing technique.
Conversion processing means for converting the pixel distribution of the input image into a pixel value so that the feature amount approximates the target output value, and outputting the result.

  First, the region-of-interest setting means sets a region of interest for the input image based on imaging part information corresponding to the imaging technique. In other words, the region of interest can be reliably identified from the input image by this setting means. For this reason, input images having common imaging region information can be extracted feature amounts described later with respect to almost common regions of interest.

  As the input image, an X-ray generator can be used that irradiates the subject with X-rays and the transmitted X-rays of the subject are detected as X-ray images mainly by the X-ray detection unit. Typically, preprocessing such as offset correction processing, gain correction processing, scratch correction processing, and black and white inversion of the image is performed on the X-ray image, and this preprocessed image is used as an input image. In addition, the X-ray image can be used as an input image by removing an area (direct line area) where X-rays reach the X-ray detection unit directly without passing through the subject from the X-ray image (pre-processed image). good. In addition, when the X-ray generation field is equipped with an X-ray diaphragm that defines the X-ray irradiation field, it is limited to the irradiation field from which the X-ray diaphragm area has been removed from the X-ray image (pre-processed image). The input image may be specified. An IP (imaging plate), an image intensifier, an FPD (planar detector), etc. having a two-dimensional position resolution can be suitably used for the X-ray detection unit.

  A region of interest is set for such an input image. This setting is performed based on imaging part information corresponding to the imaging technique. At the time of shooting, first, a shooting technique is selected. Therefore, if a region of interest corresponding to the procedure ID (chest front, abdomen, head, etc.) is set in advance, the region of interest can be set for the input image. For example, in the photographed image of the front of the chest, since the outline of the photographing body position is known from the photographing technique and technique, the rough position of the lung field can be estimated. Therefore, the center coordinates and size of the region of interest may be set based on the approximate position of the lung field in the input image.

  The number of regions of interest may be set as appropriate according to the imaging region, and may be singular or plural. For example, a chest front image that requires compression of the dynamic range is composed of an image of a lung field that easily transmits X-rays and an image of a mediastinum that does not easily transmit X-rays. It is preferable to set the region of interest in two places of the part. In the case of an image of the head, the subject is usually located in the center of the image, and the degree of X-ray transmission with respect to the subject does not vary greatly, so if one region of interest is set in the center of the input image. good.

  The shape of the region of interest may be determined as appropriate according to the shape of the subject (part) so as to cover the specific region of the subject. For example, a rectangular region of interest is preferable for the lung field, and a circular region of interest is preferable for the head.

  The shape of the region of interest does not need to be accurately matched with the shape of the subject part, and may be an appropriate shape that can cover most of the target subject part. For example, when the subject region is the lung field, it is not necessary to accurately extract the shape of the lung field and set the region of interest according to the extracted shape, and the rectangular region of interest may be set as described above.

  However, when the imaging region is an extremity, it is preferable that the entire effective diagnosis region excluding the direct line region from the irradiation field is the region of interest. In this case, since the effective diagnosis region is substantially the contour shape of the extremity to be imaged, the region of interest having this contour shape is set.

  The size of the region of interest may be fixed or variable. By changing the size of the region of interest, the region of interest can be set according to the size of the subject. In order to change the size of the region of interest, for example, a direct line region is excluded from the input image, and a region where the pixel value is about 20% to 80% of the maximum value may be set as the region of interest.

  Next, the feature amount extraction unit extracts the feature amount of the pixel value from the region of interest described above. As described above, since the feature amount is always extracted from the set region of interest, the feature amount is not extracted from an unrelated region in interpretation.

  This feature amount is set in advance together with the region of interest in accordance with the imaging region. For example, in the case of an input image in front of the chest, as described above, the region of interest is set in the lung field and the mediastinum. In this case, since the lung field easily transmits X-rays, the maximum value of the pixel value is set as a feature amount, and the mediastinum portion is difficult to transmit X-rays. Therefore, the minimum value of the pixel value may be set as the feature amount. Furthermore, in addition to the feature amount extracted from the region of interest, a feature amount extracted from another region may be used together. For example, an average value of pixel values of the entire input image can also be used as the feature amount.

  In addition, local measurement values of various physical quantities such as X-ray dose, optical density, luminance, X-ray absorption coefficient, hydrogen atom concentration, radioisotope concentration, ultrasonic reflection amount, and temperature can be used as pixel values. it can.

  Next, the target output value setting means sets a target output value corresponding to the feature amount for each photographing technique. This target output value is stored in the storage means as a value that makes the gradation characteristics of the converted and output image appropriate when the pixel value conversion of the feature value is performed by the conversion processing means described later. . For example, the target output value may be set so that the feature amount of the lung field region has an output density of 1.8 and the feature amount of the mediastinum region has an output density of 0.3.

  Next, the conversion processing means converts the pixel distribution of the input image into pixel values so that the feature amount approximates the target output value, and outputs the result. By this pixel value conversion, the gradation characteristics of the converted and output image can be made appropriate, and a high-quality image can be obtained over the entire area of the photographed image. More specifically, for example, the pixel distribution of the input image is smoothed to obtain a smoothed pixel distribution, and the smoothed pixel distribution is subjected to pixel value conversion and output. That is, the dynamic range of the low-frequency component image obtained by the smoothing process is compressed by this smoothing process and pixel value conversion. In this smoothing process, for example, the input image is smoothed with an average filter having an appropriate mask size. By this smoothing process, an image obtained by removing high frequency components from the input image is obtained.

  Next, the pixel value of each pixel value of the input image, for example, the smoothed image is converted. At that time, the conversion is performed so that the previously obtained feature value approximates the target output value. Typically, an appropriate curve is generated so that the feature amount approaches the target output value, and pixel value conversion is performed using the appropriate curve. The appropriate curve is generated by, for example, “feature coordinates and origin”, “multiple feature coordinates”, “feature coordinates” in finite coordinates with the input pixel value as the horizontal axis and the output pixel value as the vertical axis. Find the slope of a straight line that passes through the "and end point". Subsequently, a multi-degree polynomial is obtained which is the inclination for which the coordinates of each feature amount and the inclination at the end point are obtained. Then, the pixel value may be converted and output using a curve represented by a multi-degree polynomial as an appropriate curve.

  The image after the pixel value conversion as described above is an image using a component image with a low spatial frequency, and the output image after the final DRC processing is usually added to the high frequency component image and output. To do. The high frequency component image is obtained by subtracting the pixel value of the smoothed image from the input image. This high frequency component image is preferably multiplied by an enhancement coefficient corresponding to the low frequency component image and then added to the low frequency component image after the pixel value conversion processing. Thereby, it is possible to obtain an output image with more emphasized contours and the like.

  According to the apparatus of the present invention, since the region of interest is set by the region of interest setting means based on the imaging region information corresponding to the imaging technique, the region of interest can be reliably identified from the input image. For this reason, input images with common imaging region information can be extracted feature amounts with respect to almost common regions of interest, and variations in the density of processed images can be suppressed.

  In addition, since the feature amount extraction unit always extracts the feature amount from the set region of interest by the feature amount extracting unit, the feature amount extraction unit does not extract the feature amount from an unrelated region in interpretation. Therefore, it is possible to always display an area necessary for interpretation with high quality contrast.

  Furthermore, the apparatus of the present invention sets a target output value corresponding to the feature amount for each photographing technique by the target output value setting means. Then, the conversion processing means converts the pixel distribution of the input image into pixel values so that the feature amount approximates the target output value, and outputs the result. As a result, a clear image with high contrast can be obtained in all regions, particularly the region of interest.

  Embodiments of the present invention will be described below with reference to the drawings.

<Example 1>
First, the following embodiment will be described by taking as an example a case where dynamic range compression is performed on an X-ray image of chest front radiographing.

  As shown in FIG. 1, the apparatus of the present invention includes an X-ray generator 101 and a two-dimensional X-ray sensor 102, and acquires an X-ray image of a subject P using these. The acquired image is supplied to the image processing apparatus 200 via the data collection circuit 103 and the preprocessing circuit 104. The image processing apparatus 200 performs image processing including dynamic range compression conversion on the supplied image. A series of these operations in the apparatus of the present invention is performed using the operation panel 105, and the control of the apparatus and data transmission associated therewith are performed using the CPU 106, the main memory 107, and the CPU bus 108. Then, the image-processed image is output to the image display circuit 109 and displayed on an appropriate display (not shown) such as a monitor via the display circuit 109.

  Hereinafter, the configuration of each unit will be described in more detail.

  The X-ray generator 101 irradiates the subject P with an X-ray beam. For example, the generator 101 has an X-ray tube that generates X-rays when a high voltage is applied. Furthermore, it is good also as a structure which has X-ray aperture_diaphragm | restriction which controls the irradiation field of X-ray | X_line.

  The two-dimensional X-ray sensor 102 receives the X-ray transmitted through the subject P while being attenuated, and outputs an X-ray image. The X-ray image of the front of the chest output here includes an image of a lung field through which X-rays are easily transmitted and a mediastinal portion through which X-rays are very difficult to transmit.

  The data collection circuit 103 collects a captured image output from the two-dimensional sensor 102. More specifically, the data acquisition circuit 103 converts the X-ray image output from the two-dimensional X-ray sensor 102 into an electrical signal and supplies it to the preprocessing circuit 104.

  The preprocessing circuit 104 performs preprocessing such as offset correction processing, gain correction processing, scratch correction processing, and black and white inversion of an image. The preprocessed X-ray image signal is transferred as an input image to the main memory 107 and the image processing apparatus 200 via the CPU bus 108 under the control of the CPU 106.

  The operation panel 105 is used for instructing execution of X-ray imaging and setting imaging conditions for each imaging region to the apparatus. For example, a touch panel display or a mechanical switch is used for the operation panel.

  The CPU 106 controls the operation of the entire apparatus of the present invention. The main memory 107 stores data used for the control of the device of the present invention and various processes, and stores various setting conditions, and the CPU bus 108 transmits signals and signals between the components constituting the device of the present invention. Used for data exchange.

  The image display circuit 109 performs predetermined signal processing for displaying a desired image on a display (not shown).

  On the other hand, the image processing apparatus 200 includes an irradiation region recognition circuit 201, a direct line removal circuit 202, a ROI (Region of interest) analysis circuit 203, a feature amount extraction circuit 204, a target output value setting circuit 205, and a DRC circuit 206. Hereinafter, the configuration and processing operation of the image processing apparatus will be described in detail.

First, when an X-ray stop is used in X-ray imaging, the irradiation region recognition circuit 201 removes the X-ray stop region from the X-ray image and identifies the irradiation field. The irradiation field is identified by obtaining the geometric conditions from the distance between the focal point of the X-ray generator 101 and the two-dimensional X-ray sensor 102 and the position information of the X-ray diaphragm, and calculating the irradiation area on the two-dimensional X-ray sensor 102. Can be used. Since the distance between the X-ray focal point and the two-dimensional sensor 102 and the position information of the X-ray diaphragm are set on the operation panel 105, they are sent to the irradiation area recognition circuit 201 via the CPU bus 108 under the control of the CPU 106. Specifically, the irradiation width W of the X-ray on the two-dimensional X-ray sensor 102 is X-ray radiation when the X-ray is irradiated when the distance between the X-ray focal point and the two-dimensional X-ray sensor 102 is D. The angle a is calculated by the following formula 1.1.
W = 2 x D x tan (a / 2) (Formula 1.1)

  Next, the direct line removal circuit 202 deletes a direct line area (area where X-rays are directly irradiated on the sensor surface) from the X-ray image. As a method for extracting the direct line area, two-dimensional X in the case where X-rays are directly irradiated on the sensor surface from the imaging information of the apparatus of the present invention, that is, the tube voltage, tube current, imaging time, additional filter, grid type, imaging distance. Since the pixel value of the line sensor 102 can be calculated, a region exceeding this pixel value can be directly obtained as a line region.

  By these image analyses, a diagnostic effective region is obtained by deleting the outside of the irradiation region (X-ray aperture) and the straight line region.

  Next, a region of interest (ROI) is set by the ROI analysis circuit 203 for the purpose of extracting the lung field portion and the mediastinum portion for this effective diagnosis region. At the time of shooting, the shooting technique is selected first, so the ROI setting and feature amount corresponding to the technique ID (front of the chest) are set in advance. In the case of a photographed image of the front of the chest, since the outline of the photographing body position is known from the photographing technique and technique, the approximate position of the lung field can be estimated, and the center coordinates and size of these feature extraction regions are set.

  Specifically, as shown in FIG. 2, two rectangular ROIs (region A) are provided symmetrically with respect to the vertical center line of the image. This rectangular ROI is set to a size that almost covers the left and right lung fields. In the extraction of the mediastinum part, one circular ROI (region C) whose center is located on the vertical center line is set at the lower part of the image excluding the ROI of the lung field. Furthermore, the entire diagnosis effective region is set as one of the ROIs (region B). That is, in the ROI analysis circuit of this example, three types of ROIs are set for the left and right lung fields (region A), the mediastinum (region C), and the entire diagnosis effective region (region B). In FIG. 2, all ROIs are indicated by broken lines.

  The feature amount extraction circuit 204 (FIG. 1) extracts feature amounts for the set ROI. That is, the minimum value, average value, maximum value, etc. of the pixel value are calculated for the set ROI. For example, if the imaging region is the front of the chest, the average pixel value is calculated from the entire image (region B), the maximum pixel value is calculated from the left and right lung fields (region A), and the mediastinum (region C) ) To calculate the minimum pixel value.

  For X-ray chest images, dynamic range compression is required in areas where the mediastinum is always low in density and in areas where the density in the center of the lung field is very high. Since the minimum value of the pixel value in the rectangular region of interest (in the case of a bone white image or a reverse image) is the feature amount of the lung field region, it is not always necessary to extract the exact shape of the lung field region. In addition, since the maximum pixel value in the circular region of interest (in the case of a bone white image or a reverse image) is the feature amount of the mediastinum (including the spine), it is not necessary to extract the exact shape of the mediastinum. is there.

  After the feature amount extraction is performed by the feature amount extraction circuit 204 as described above, the DRC circuit 206 performs dynamic range compression processing on the input image. The configuration and function of the DRC circuit 206 will be described with reference to the functional block diagram of FIG. The DRC circuit 206 includes an average filter 206A, a compression LUT (Look up table) 206B, an enhancement coefficient LUT (Look up table) 206C, and arithmetic means for performing arithmetic processing of various pixel values.

  First, an input image that has been subjected to direct line removal processing with respect to an irradiation area is smoothed by an average filter 206A having a mask size of about 101 × 101. The high frequency component is removed by this smoothing, and the obtained low frequency component image M is compressed by the compression LUT 206B.

  The compression LUT 206B corresponds to the gradation characteristics of low frequency components, generates an appropriate curve for gradation conversion for the input image, and converts and outputs the input pixel value based on this curve. More specifically, the relationship between the two pixel values is defined with the input pixel value as the horizontal axis and the output pixel value as the vertical axis so that the previously determined feature value of the front of the chest approaches the target output value (target density). Generate an appropriate curve. The method for generating the appropriate curve will be described in detail later.

  As the target output value, a value stored for each imaging region in the target output value setting circuit 205 (FIG. 1) is used. For example, the target output value is set so that the feature amount of the lung field region has an output optical density of 1.8 on the vertical axis and the feature amount of the mediastinum region has an output optical density of 0.3 on the vertical axis.

Then, by subtracting the low frequency component image M (FIG. 3) from the input image, and extracts a high frequency component V _0. At the same time, the enhancement coefficient β is determined according to the low frequency component image M by the enhancement coefficient LUT206C. Then, the DRC-processed image is output by adding the high-frequency component V multiplied by the enhancement coefficient β to the low-frequency component M ′ after applying the compression LUT.

  Next, a method of generating a compression LUT curve for a 12-bit image (4096 gradations) will be described with reference to FIG. Here, using the three types of feature values of the maximum value X1 of the right lung field and the left lung field, the average value X2 of the entire diagnosis effective area, and the minimum value X3 of the mediastinum, the processed image after DRC processing for the maximum value X1 A curve is generated with the target pixel value Y1, the target pixel value of the processed image after DRC processing for the average value X2 is Y2, and the target pixel value of the processed image after DRC processing for the minimum value X3 is Y3.

  (1) The control points are the origin (0,0), (x1, y1), (x2, y2), (x3, y3), and the end point (4095,4095).

(2) Calculate the slope a ₁ at x = x ₁ . The slope a ₁ is the slope of a straight line passing through the points (x ₁ , y ₁ ) and (x ₂ , y ₂ ), and is represented by the following formula 2.1.

(3) Calculate the slope a ₂ at x = 4095. The slope a ₂ is the slope of a straight line passing through the points (x ₃ , y ₃ ) and (4095, 4095), and is represented by the following formula 2.2.

(4) point (x _1, y _1), of passing the (4095,4095), cubic polynomial slope at x = x ₁ is the slope of a _1, x = 4095 is a ₂ (formula 2.3) Each coefficient (a _H , b _H , c _H , d _H ) is calculated.
y = a _H x ³ + b _H x ² + c _H x + d _H x: [x ₁ , 4095] (Formula 2.3)

(5) x: [0, x ₁ −1], passing through points (0, 0) and (x ₁ , y ₁ ), and a quadratic polynomial (following Equation 2.4 below) having a slope a _{1 at} x = x ₁ ) Of each coefficient (a _L , b _L ). Each coefficient (a _L , b _L ) is obtained by the following formulas 2.5 and 2.6.
y = a _L x ² + b _L xx: [0, x ₁ −1] (Formula 2.4)

  (6) The compression LUT of the DRC circuit 206 is created as a curve represented by equations 2.3 and 2.4.

  As described above, according to the apparatus of the present invention, the dynamic range compression processing is executed depending on the feature amount of the region of interest set for each imaging region. Thereby, it is possible to efficiently obtain a high-quality and stable image at all times.

  Also, the compression LUT curve when performing dynamic range compression can be automatically determined depending on the characteristics of the region of interest corresponding to the imaging region. For example, since the compression LUT curve is determined so as to pass through the minimum value of the mediastinum and the maximum value of the lung field, the density of the region of interest is kept almost constant in the image after compression conversion output, and the blood vessel Is kept constant regardless of the structure of the subject. Therefore, an X-ray chest image that is always stable can be obtained.

  Furthermore, by using the information of the imaging technique, the approximate position of the image-like region of interest can be known, and a high-quality image can be obtained with sufficient accuracy without performing detailed analysis of the region of interest.

  In the above description, only the chest front image has been described. However, the ROI and the feature amount corresponding to the technique ID can be obtained in the same manner for other regions.

<Example 2>
Next, the embodiment of the present invention will be described taking the case where the imaging region is the front of the abdomen as an example. However, in the case of this example as well, the configuration and functions of the apparatus used are the same as those in the first embodiment, and here, the description will focus on differences in the processing procedure of the image processing based on the difference in the imaging region.

  In the case of abdominal frontal imaging, one circular region of interest (region D) may be set at the center of the image as shown in FIG. The circular size of the region of interest may be variable. If the minimum value and maximum value of the entire diagnostic effective area in the image and the average value of the circular area at the center of the image are used as the feature amount, a compression LUT curve can be generated as in the first embodiment. The target output value may be set so that the average value of the circular area in the center of the image is 50% of the output range (for example, 4095 × 0.5 = 2048).

<Example 3>
Next, the embodiment of the present invention will be described by taking the case where the imaging region is the head side surface as an example. Also in this example, since the configuration and functions of the apparatus used are the same as those in the first embodiment, the description will focus on differences in the processing procedure of the image processing based on the difference in imaging region.

  In the case of head imaging, as shown in FIG. 6, since the subject comes to the center of the X-ray cone as an imaging technique, one circular region of interest (region E) is provided at the center of the X-ray image. The size of the region of interest is variable so that it does not protrude directly into the line region. The feature amount may be the minimum value, the maximum value, and the average value of the pixel values in this region of interest. The target output value may be set so that the average converted pixel value is 30% of the output range (eg, 4095 × 0.3 = 1228) in consideration of neck extraction. In FIG. 6, the intersection of the vertical center axis and the horizontal center axis is the center of the X-ray cone, the rectangular broken line area is the irradiation field, and the black rectangular area is the direct line area.

<Example 4>
Next, the embodiment of the present invention will be described by taking the case where the imaging region is the forearm as an example. Also in this example, since the configuration and functions of the apparatus used are the same as those in the first embodiment, the description will focus on differences in the processing procedure of the image processing based on the difference in imaging region.

  In the case of limb photography, as shown in FIG. 7, the subject may not come to the center of the X-ray cone. Further, there are cases where the observation region is wide and the region of interest is the peripheral portion of the image. However, since the limb image has a relatively narrow dynamic range of X-rays, the compression LUT curve may be simple. For example, the entire diagnostic effective area obtained by excluding the direct line area (black area) from the irradiation field excluding the X-ray diaphragm may be set as the area of interest (hatched area F). The feature amount is an average value, a minimum value, and a maximum value of the pixel values of the entire diagnosis effective area. Then, the target output value may be set so that the converted pixel value of the average value is 50% of the output range (for example, 4095 × 0.5 = 2048).

  The apparatus of the present invention can be suitably used for image processing of medical X-ray images.

1 is a functional block diagram of a device of the present invention related to Example 1. FIG. FIG. 6 is an explanatory diagram illustrating a procedure for setting a region of interest in an image in front of the chest using the apparatus according to the first embodiment. It is a functional block diagram of a DRC circuit in the device of the present invention. It is explanatory drawing which shows the example of a production | generation of the compression LUT in a DRC circuit. FIG. 10 is an explanatory diagram illustrating a procedure for setting a region of interest in an image of the front of the abdomen in the second embodiment. FIG. 10 is an explanatory diagram illustrating a procedure for setting a region of interest in an image of a head side surface in the third embodiment. FIG. 10 is an explanatory diagram illustrating a procedure for setting a region of interest in an image of a forearm in Example 4.

Explanation of symbols

101 X-ray generator
102 2D X-ray sensor
103 Data acquisition circuit
104 Pre-processing circuit
105 Operation panel
106 CPU
107 Main memory
108 CPU bus
109 Image display circuit
200 Image processing device
201 Irradiation area recognition circuit 202 Direct line removal circuit 203 ROI analysis circuit
204 Feature extraction circuit 205 Target output value setting circuit 206 DRC circuit
206A Average filter 206B Compression LUT 206C Enhancement factor LUT
P Subject A, B, C, D region (region of interest)

Claims (2)

A medical image processing apparatus for compressing a dynamic range of an input image obtained by photographing,
A region-of-interest setting means for setting a region of interest based on imaging part information corresponding to an imaging technique for the input image;
Feature quantity extraction means for extracting feature quantities of pixel values constituting the input image from the set region of interest;
Target output value setting means for setting a target output value corresponding to the feature amount for each shooting technique;
A medical image processing apparatus, comprising: conversion processing means for converting the pixel distribution of the input image so as to approximate the target output value to the target output value, and outputting the converted pixel value. Furthermore, it has a smoothing means for smoothing the pixel distribution of the input image,
The medical image processing apparatus according to claim 1, wherein the conversion processing unit performs pixel value conversion on the pixel distribution smoothed by the smoothing processing unit and outputs the result.

Images