JP2010226634A - Image defect detection method and radiographic imaging apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a defective pixel detection method and a radiographic imaging apparatus in which a defective pixel of a radiograph detector is properly detected and excessive correction can be suppressed. <P>SOLUTION: A first threshold is used to detect a temporary defect from a detection image of a defective pixel, detected defective pixels are classified into long defects and other defects, and a defective pixel is re-detected from the detection image by making a second threshold, which is equal to or greater than the first threshold, corresponding to the long defects and making a third threshold, which is greater than the first threshold, corresponding to the other defects. Regarding the defective pixel detected with the second threshold, information indicating that correction is not performed on a defect of a length exceeding a predetermined threshold is added. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an image defect detection method and a radiographic imaging apparatus suitable for a breast cancer inspection apparatus, and more specifically, a defective pixel detection method capable of appropriately correcting only defective pixels that need correction, and the defect The present invention relates to a radiographic imaging apparatus using a pixel detection method.

  When breast cancer screening is performed, the detection rate of early cancer increases when a breast cancer inspection device (mammography) that captures a radiographic image of a breast is combined with screening by visual inspection alone. Therefore, in breast cancer screening, in addition to (or instead of) visual inspection, screening using a breast cancer inspection apparatus is performed.

  In breast cancer inspection equipment, with the breast placed on the imaging table that contains the radiation image detector, the breast is pressed by the compression plate, irradiated to the breast from the compression plate side, and the radiation that has passed through the breast is radiated. A radiation image of the breast is taken by receiving light with the image detector.

  The radiographic image detector used in various radiographic imaging apparatuses such as an X-ray diagnostic apparatus as well as such a breast cancer inspection apparatus is a radiation (X-ray, α-ray, β-ray, γ-ray) transmitted through a subject. A radiation image is taken by taking out an electric signal such as an electron beam, an electron beam, or an ultraviolet ray.

As one type of radiation image detector, a flat radiation image detector, a so-called flat panel detector (hereinafter referred to as FPD) is known.
Further, as such an FPD, for example, electron-hole pairs (e-h pairs) emitted from a photoconductive film such as amorphous selenium by the incidence of radiation are collected and read out as electrification signals. It has a direct method for converting it into an electrical signal and a phosphor layer (scintillator layer) formed of a phosphor that emits light (fluoresce) when incident on radiation. This phosphor layer converts radiation into visible light, There are two methods of reading light with a photoelectric conversion element, that is, an indirect method of converting radiation into an electric signal as visible light.

By the way, in a radiographic imaging apparatus using FPD, defective pixels of FPD are cited as a cause of deterioration in image quality of radiographic images.
That is, not all FPD pixels (detection elements) necessarily output a signal having an appropriate intensity with respect to incident radiation (radiation dose). There are pixels that output a signal with a reasonably high value.

As a matter of course, an appropriate radiation image cannot be obtained in a portion (pixel) where a defective pixel is generated. An image having such a defect causes a serious problem such as misdiagnosis. Further, the number of defective pixels of FPD cannot be prevented from increasing with time.
For this reason, in a radiographic image capturing apparatus using FPD, it is normal to perform defective pixel correction and reproduce the defective pixel corrected radiographic image as a diagnostic image or the like as an image display or print. As an example, this defective pixel correction detects the position of the defective pixel of the FPD at a predetermined timing, and when photographing / outputting a radiographic image, according to the detection result of the defective pixel, the peripheral pixel (the image) Data) is used to correct defective pixels.

For example, in Patent Document 1, when a defective pixel of an FPD is detected, the defective pixel is classified according to a pattern according to the number of consecutive defective pixels, and a correction coefficient according to the pattern is applied to be adjacent to the defective pixel. FPD defective pixel correction is described in which defective pixels are corrected by interpolation using image signals of pixels (excluding defective pixels).
In Patent Document 2, a defective pixel of an FPD is detected using first image data, and a defective pixel detected from the first image data with respect to second image data obtained by capturing a radiation image. Radiation image processing apparatus that generates third image data in which defective pixels are corrected in accordance with information, and displays the second image data and the third image data at the same time, switching display, composite display, etc. Is described.

Japanese Patent Laid-Open No. 10-51693 JP 2000-126162 A

However, in such a conventional FPD defective pixel correction method, not only defective pixels that need to be corrected, but also defective pixels that do not require correction, such as defective pixels that are less likely to cause misdiagnosis, and conversely, correction is performed. If so, there is a case where so-called overcorrection that corrects a very large collection of defective pixels, which may cause misdiagnosis, may be performed.
On the contrary, such overcorrection may cause misdiagnosis, and improvement is desired.

  An object of the present invention is to solve the above-mentioned problems of the prior art, and it is possible to properly detect defective pixels of an FPD used in a radiographic imaging apparatus, and to correct defective pixels that do not require correction. An object of the present invention is to provide an FPD defective pixel detection method capable of suppressing correction and preventing overcorrection, and a radiographic imaging apparatus using this detection method.

  In order to achieve the above object, the defective pixel detection method according to the present invention detects a defective pixel of a radiation image detector in which pixels are two-dimensionally arranged in the xy direction. A step of acquiring a detection image for detecting a pixel, a step of detecting a provisional defective pixel using a first detection threshold by using the detection image, and at least one of the x direction and the y direction. For the provisional defective pixel, “Δx / Δy”, which is the ratio of the number of pixels Δx continuous in the x direction and the number of pixels Δy continuous in the y direction, is calculated. A temporary defect other than the long defect is detected by detecting a long defect that is equal to or less than a first selection threshold or equal to or greater than a second selection threshold and the second detection threshold that is equal to or greater than the first detection threshold. The first detection threshold for the defective pixels of Detecting a defective pixel of the radiation image detector using the detected image with a third detection threshold value greater than the first detection threshold; generating information indicating a position of the defective pixel on the radiation image detector; 2 For the defective pixels detected by the detection threshold, the number of pixels continuous in the x direction and the y direction is detected, and the number of continuous pixels in the larger x direction and y direction is determined to be a defective pixel that is equal to or greater than the third selection threshold. On the other hand, there is provided a method for detecting a defective pixel, characterized in that information indicating the position of the defective pixel is added with information indicating that it is a defective pixel that does not require correction.

  Further, the radiographic imaging device of the present invention is a detection source for detecting a radiation source, a radiographic image detector in which pixels are two-dimensionally arranged in the xy direction, and a defective pixel of the radiographic image detector. An image is acquired, and using this detection image, a provisional defective pixel is detected by a first detection threshold, and the provisional defective pixel continuous in at least one of the x direction and the y direction is detected in the x direction. “Δx / Δy”, which is a ratio between the number of consecutive pixels Δx and the number of consecutive pixels Δy in the y direction, is calculated, and this “Δx / Δy” is less than or equal to the first selection threshold or greater than or equal to the second selection threshold. A certain long defect is detected, and for the long defect, the second detection threshold value is equal to or greater than the first detection threshold value, and for the temporary defect pixel other than the long defect, the first detection threshold value. Using the detected image with a third detection threshold greater than Redetecting defective pixels of the radiation image detector, generating defect position information indicating the position of the defective pixels on the radiation image detector, and for the defective pixels detected by the second detection threshold, Defects that do not require correction in the defect position information are detected for defective pixels in which the number of consecutive pixels in the x and y directions is greater than or equal to the third selection threshold. In accordance with the defect position information created by the defective pixel detection means and the defective pixel detection means that attaches information indicating that the pixel is a pixel, a defective pixel that does not require correction with respect to the radiation image captured by the radiation image detector. Provided is a radiographic imaging apparatus comprising defective pixel correction means for correcting defective pixels other than those to which information indicating that there is information.

In such a defective pixel detection method and radiographic imaging apparatus of the present invention, it is preferable that the first selection threshold is 0.8 and the second selection threshold is 1.5.
Further, two specification limits corresponding to the number of consecutive defective pixels are set, and a defective pixel whose high specification limit is detected by the third detection threshold is detected for the defective pixel detected by the second detection threshold. Preferably, a low specification limit is selected for each, and a specification limit corresponding to the defective pixel detected by the third detection threshold is set as a standard for the radiation image detector. Preferably, it is a specification limit, and it is preferable that the third selection threshold corresponds to a specification limit of defective pixels detected by the third detection threshold, and further, the number of consecutive defective pixels is a specification limit. It is preferable to output a warning when the value exceeds.
Furthermore, in the defective pixel detection method of the present invention, it is preferable to further correct a defective pixel other than the defective pixel to which information indicating that the defective pixel does not need correction is added.

According to the present invention having the above-described configuration, a defective pixel of an FPD (Flat Panel Detector radiation image detector) is detected, and the detected defective pixel has a shape of a detected defect (the shape of a continuous defective pixel). Defective pixel detection is performed again with different threshold values. Here, in the second defective pixel detection, the defect detection is performed with a small threshold value (however, more than the first detection threshold value) in a long defect region continuous in the x direction or the y direction, and in other regions. Performs defect detection with a large threshold.
A long defect is unlikely to be caused by noise or the like, that is, it is highly likely that an appropriate defective pixel is not detected erroneously. Also, such defects often have a low concentration. On the other hand, a defect (we have the same number of pixels in the x-y direction) that is axial may be noise, and the possibility of false detection cannot be denied. Therefore, according to the present invention having the above-described configuration, it is possible to reliably detect defective pixels that are continuous in a long length, and to appropriately prevent erroneous detection of noise or the like as defective pixels.

Further, among the long defects, defective pixel correction is not performed for defective pixels in which the number of consecutive pixels in the x and y directions exceeds a predetermined threshold.
The long pattern is highly likely to be caused by artifacts, and the possibility of misdiagnosis with a lime portion such as breast cancer is extremely low. In addition, if a very large defect is corrected, all information in this area will be lost, so if there is useful information such as calcification there, the correction will lead to misdiagnosis. There is a possibility.
Therefore, according to the present invention in which correction of a large continuous defective pixel is not performed, an abnormality of the FPD can be visually confirmed. Therefore, misdiagnosis caused by not correcting a defective pixel can be avoided and a large area can be detected. Misdiagnosis due to overcorrection can also be suitably suppressed.

It is a figure which shows notionally an example which utilized the radiographic imaging apparatus of this invention for the breast cancer test | inspection apparatus. It is a figure which shows notionally the imaging stand of the breast cancer test | inspection apparatus shown in FIG. It is a block diagram which shows notionally an example of the image process part of the breast cancer test | inspection apparatus shown in FIG. It is a flowchart for demonstrating an example of the defect detection method of this invention. It is a figure which shows notionally an example of the defective pixel for demonstrating the defect detection method of this invention.

  Hereinafter, a defective pixel detection method and a radiographic imaging apparatus according to the present invention will be described in detail based on preferred embodiments shown in the accompanying drawings.

FIG. 1 shows a conceptual diagram of an example in which the radiographic imaging apparatus of the present invention that implements the defective pixel detection method of the present invention is used in a breast cancer inspection apparatus 10.
As shown in FIG. 1, the breast cancer inspection apparatus 10 basically includes an imaging table 12, a radiation irradiation unit 14, a compression means 16, an arm 18, a base 20, and a high-voltage power supply 22 for X-ray irradiation. And an image processing unit 30 (see FIG. 3). The breast cancer inspection apparatus 10 in the illustrated example is basically a normal breast cancer inspection apparatus (mammography (radiation image of breast) except for performing defective pixel detection and defect pixel correction of the FPD 56, which is a feature of the present invention described in detail later. It is the same as the photographing apparatus)). 1 and the like, the symbol M in the figure conceptually indicates a breast, and the symbol H indicates a subject (its chest wall).

In the breast cancer inspection apparatus 10 shown in the figure, the arm 18 is substantially C-shaped bent at two right angles, and the radiation irradiation unit 14 is fixed to the upper end and the imaging table 12 is fixed to the lower end. The compression means 16 is fixed between the radiation irradiation unit 14 and the imaging table 12.
The arm 18 is supported on the base 20 by a shaft 24. Inside the base 20, a rotating means and a lifting / lowering means for the shaft 24 are incorporated. The arm 18, that is, the imaging table 12 and the radiation irradiation unit 14 are raised and lowered by raising and lowering the shaft 24 by the raising and lowering means, and rotated (rotated in a direction perpendicular to the paper surface of FIG. 1) by rotating the shaft 24 by the rotating means. It is adjusted to correspond to MLO shooting or the like.

The radiation irradiating unit 14 is provided with an operating means 26a, and the arm 18 is provided with an operating means 26b. Further, the base 20 is provided with operation means 28.
In the illustrated example, both the operation means 26a and the operation means 26b are provided with a switch for rotating and raising / lowering the arm 18, a light irradiation field lighting switch, and the like. The operation means 28 is a foot pedal type operation means connected to the base 20 by a cable 28a, and is provided with a switch for raising and lowering a compression plate 38 to be described later, a switch for raising and lowering the arm 18, and the like.

  The radiation irradiation unit 14 is a part that irradiates the breast M (FPD 56) with radiation, and, like a known breast cancer inspection apparatus, a radiation source, a collimator that regulates the radiation field, and a drive unit that drives the radiation source And so on.

The compression means 16 compresses the breast M against the imaging table 12 during imaging, and includes a compression plate 48 that compresses the breast M against the imaging table 12 and an elevating means 50 for the compression plate 48. The compression plate 48 is configured to be detachable from the elevating means 50. As an example, an object of 18 × 24 cm size corresponding to a normal size breast M and a size of 24 × 30 cm size corresponding to a large breast M are provided. It is prepared.
In the illustrated breast cancer testing apparatus 10, the compression plate 48 and the lifting / lowering means 50 are basically a known compression plate and its lifting / lowering means provided in a known breast cancer testing apparatus.

The imaging table 12 is a hollow housing whose upper surface is a mounting surface 12a of the breast M, and as shown schematically in FIG. 2, a scatter removal grid 54 and an FPD 56 are arranged inside.
Although not shown, the imaging table 12 includes an AEC (Automatic Exposure) for measuring radiation transmitted through the breast M in pre-irradiation performed prior to radiographic imaging in order to determine imaging conditions. Various members of a known breast cancer inspection apparatus such as a control sensor and moving means for the scatter removal grid 54 are appropriately arranged.

  The scatter removal grid 54 is a known grid that is disposed in the radiographic imaging apparatus in order to prevent scattered radiation from entering the FPD 56.

  The FPD 56 is a known radiation image detector (radiation solid state detector) that detects radiation irradiated by the radiation irradiation unit 14 (ray source) and transmitted through the breast M of the subject H.

In the present invention, the FPD 56 is used for various radiographic imaging devices in which pixels that detect radiation two-dimensionally are arranged in the xy direction (x direction and y direction orthogonal to the x direction). A known FPD (Flat Panel Detector).
Therefore, in the present invention, various FPDs 56 can be used. That is, it has a photoconductive film such as amorphous selenium, collects charges (electron-hole pairs (e-h pairs)) generated by the photoconductive film upon incidence of radiation, and reads them as electric signals. Even in FPD, a scintillator layer formed of a phosphor that emits light (fluorescence) such as “CsI: Tl” and a photodiode are used, and light emission of the scintillator layer due to the incidence of radiation is photoelectrically converted by the photodiode. Alternatively, a so-called indirect FPD that reads out as an electrical signal may be used.

A radiation image of the breast M (an output signal of the FPD 56) captured by the FPD 56 is output to the image processing unit 30.
The image processing unit 30 processes the output signal output from the FPD 56 and displays an image (image data (image data (image data) corresponding to display on the monitor 32, print output on the printer 34, and output using a network or a recording medium). Signal)). In the photographing apparatus 10 of the illustrated example, the image processing unit 30 includes a data processing unit 60, a defect detection unit 62, a defect correction unit 64, and an image processing unit 68 as conceptually shown in FIG.

Such an image processing unit 30 includes, as an example, one or a plurality of computers and workstations. In addition to the illustrated parts, various operations, input of instructions, and the like are necessary. Has a keyboard, mouse, etc.
Further, the image processing unit 30 (such as a computer constituting the image processing unit 30) controls and manages the entire breast cancer inspection apparatus 10, and controls and manages the operation of the breast cancer inspection apparatus 10. Control means 70 is provided.

  The data processing means 60 performs predetermined processing such as A / D conversion and log conversion on the output signal of the FPD 56 to convert it into a radiation image of the breast M (its image data (image signal)), and the defect detection means 62. And supplied to the defect correcting means 64.

The defect detection means 62 is a characteristic part of the present invention, and processes the detection image (radiation image for defect detection (image data)) supplied from the data processing means 60 to detect defective pixels of the FPD 56. It is to detect.
The defect correction unit 64 performs defect pixel correction of the FPD 56 on the radiation image processed by the data processing unit 60 in accordance with the detection result of the defective pixel by the defect detection unit 62.
The defect detection means 62 and the defect correction means 64 will be described in detail later.

  The image processing unit 68 performs predetermined image processing by performing predetermined image processing on the radiographic image subjected to defect pixel correction by the defect correction unit 64, displays an image on the monitor 32, and prints (hardware) by the printer 34. Output as a radiation image (its image data) corresponding to output to a network or a storage medium, and the like, to a monitor 32, a printer 34, and a designated part such as a network.

  The image processing performed by the image processing unit 68 is not particularly limited. Accordingly, the image processing means 68 performs various kinds of processing such as gain correction (shading correction), offset correction, afterimage correction, gradation correction, density correction, and data conversion for converting a radiation image into an output image such as a monitor display or print output. All of the image processing performed in the breast cancer inspection apparatus, and also in the radiographic imaging apparatus and the image processing apparatus can be used. All of these corrections may be performed by a known method.

As described above, the image processing unit 68 performs image processing on the radiation image that has been subjected to the defective pixel correction by the defect correcting unit 64. In addition, the defect correction unit 64 performs defect pixel correction on the radiation image using the information on the defective pixel detected by the defect detection unit 62.
Hereinafter, the present invention will be described in more detail by describing the operation of the defect detection means 62 and the defect correction means 64 with reference to the flowchart shown in FIG.

As is well known, a defective pixel (pixel defect) is a pixel (radiation detection element) that outputs an inappropriately high output signal or an inappropriately low output signal with respect to an incident radiation dose.
The defect detection means 62 detects a pixel defect of the FPD 30 using the detection image, and creates a defective pixel map indicating the position of the pixel defect on the FPD 30 (the position of the pixel having the defect).

  In the breast cancer inspection apparatus 10, the defect detection unit 62 acquires a detection image for detecting defective pixels of the FPD 56 at a predetermined timing. That is, in the breast cancer inspection apparatus 10, an image to be an inspection image is taken by the FPD 56 at a predetermined timing, the predetermined processing is performed by the data processing unit 56, and is supplied to the defect detection unit 62.

There are no particular limitations on the detection image, and any image can be used as long as it can detect defective pixels of the FPD 56, such as an image used for detection of defective pixels in Patent Document 1 and Patent Document 2. Is possible.
Specifically, a dark image (an image obtained by reading the dark current of the FPD 56), a radiation image obtained by uniformly irradiating the entire surface of the FPD 56 with a predetermined amount of radiation without the subject H (so-called solid image) Examples include a radiation image obtained by subtracting a dark image from a solid image. In addition, a radiation image obtained by smoothing the original image with a low-pass filter or averaging and subtracting the averaged image from the original image is also used as the detection image. It can be suitably used.

  In addition, in the breast cancer inspection apparatus 10 shown in the figure, when image processing is required for acquisition (generation) of the detection image, the image processing may be performed by either the data processing unit 60 or the defect detection unit 62. Or you may carry out in another site | part.

  Further, the acquisition timing of the detection image, that is, the creation (update) timing of the defective pixel map is not particularly limited, and after starting the breast cancer inspection apparatus 10, every predetermined number of imaging, every lapse of a predetermined time, radiation source Each time 22 irradiates a predetermined dose of radiation, it may be the same as the timing used in various breast cancer inspection devices (radiation imaging devices) such as a combination thereof.

When acquiring the detection image, the defect detection means 62 first detects a temporary defective pixel of the FPD 56 from the inspection image using the first detection threshold (1st detection Th).
The temporary defective pixel detection method using the first detection threshold may be performed in the same manner as the known defective pixel detection method of the FPD 56, such as the method described in Patent Document 1 or Patent Document 2. For example, the first detection threshold (density value (density data)) is compared with the density (absolute value) of each pixel of the detection image, and a pixel whose absolute value exceeds the first detection threshold is determined as a temporary defect. A method of detecting as a pixel is exemplified. Alternatively, a method is also used in which a reference value (median value, average value, etc.) of the detection image is determined, and a pixel whose difference from the reference value exceeds the first detection threshold in the detection image is detected as a temporary defective pixel. Is possible.

The first detection threshold value is not particularly limited, and may be appropriately determined according to the characteristics of the FPD 56, the performance required for the breast cancer inspection apparatus 10, and the like. Further, it may be determined by performing experiments, simulations, or the like, or may be determined by taking into consideration the characteristics of the FPD 56 and the like in the experiments. Furthermore, the first detection threshold value may be reset (updated) as appropriate according to time, the number of shots, the installation environment of the breast cancer inspection apparatus, and the like.
Alternatively, a temporary defective pixel may be detected from the detection image using the output signal intensity or luminance of the FPD 56 instead of the density.
The same applies to the second detection threshold and the third detection threshold described later. However, the first detection threshold is preferably set to a lower threshold so that all defective pixels can be detected without leaking.

As described above, the FPD 56 is a radiation image detector having a two-dimensional pixel array in which pixels for detecting radiation are arrayed in the xy direction.
When the defect detection means 62 detects the provisional detection pixel in this way, the defect detection means 62 detects the provisional defect pixel continuous in the x direction and / or the y direction, and determines the number of pixels (length) Δx continuous in the x direction. The number of pixels Δy continuous in the y direction is detected. In other words, a temporary defective pixel continuous in the x direction and / or the y direction is regarded as one defect, and the number of pixels Δx in the x direction and the number of pixels Δy in the y direction are detected (hereinafter referred to as x direction and / or / Alternatively, (temporary) defective pixels continuous in the y direction are collectively referred to as “defects”.
For example, if the defect is as shown in FIG. 5, the number of pixels Δx in the x direction is 4, and the length Δy of the pixel in the y direction is 6.

Next, Δx / Δy is calculated for each defect, and Δx / Δy is 0.8 or less (that is, a defect with “Δx / Δy ≦ 0.8”), and Δx / Δy is 1.5 or more. (That is, a defect having “Δx / Δy ≧ 1.5”) is detected as a long defect.
That is, in this example, as an example, a defect that is long in the x direction (lateral direction) using 0.8 as the first selection threshold and 1.5 as the second selection threshold larger than the first selection threshold. And defects that are long in the y direction (longitudinal direction) are detected as long defects, and are classified into long defects and other normal defects.
In the case of the defect shown in FIG. 5, since Δx / Δy = 4/6 = 0.67, this defect is detected as a long defect.

Next, for the long defect in the detection image, the defective pixel is re-detected using the second detection threshold (2nd detection Th) that is equal to or higher than the first detection threshold, and a temporary defective pixel other than the long defect is detected. On the other hand, a defective pixel is redetected using a third detection threshold (3rd detection Th) that is larger than the first detection threshold. However, the third detection threshold is preferably larger than the second detection threshold. That is, “first detection threshold ≦ second detection threshold”, “first detection threshold <third detection threshold”, and more preferably “second detection threshold <third detection threshold”.
In other words, in the breast cancer inspection apparatus 10, a second detection threshold that is equal to or greater than the first detection threshold and a third detection threshold that is greater than the first detection threshold are set, and the second detection threshold for a long defect. For the other defective pixels, the third detection threshold is selected, and the defective pixel is redetected from the defective pixels detected by the first detection threshold.

Note that the defective pixel detection method (redetection of defective pixels from the inspection image) using the second detection threshold and the third detection threshold is similar to the temporary detection of defective pixels using the first detection threshold. It may be performed by a known method using.
Different methods may be used for detection of a provisional defective pixel and redetection of a defective pixel.

A defect having a small Δx / Δy is a defect that is long in the y direction, and a large defect is a defect that is long in the x direction (vertical / horizontal defect).
In the detection of defective pixels of the FPD 56, a long defect, that is, a defective pixel continuous in the pixel arrangement direction, that is, the x direction or the y direction, is not caused by noise or the like, and is very likely to be a properly detected defective pixel. high. In addition, long defects often have a low density and change in density like gradation. On the other hand, a defect having the same size in the xy direction (usual defect) may be noise, and the possibility of erroneous detection cannot be denied.

Correspondingly, in the defect detection means 62, for a long defect with Δx / Δy of 0.8 or less and a long defect with Δx / Δy of 1.5 or more, the first detection threshold value or more. Using the second detection threshold, the defective pixel is re-detected from the detection image, and for other defective pixels, the third detection threshold larger than the first detection threshold is used to detect the defective pixel. Rediscover. Thereby, it is possible to suitably suppress the detection of a defect that is long in the x direction or the y direction and the erroneous detection of noise or the like as a defective pixel.
Preferably, by making the third detection threshold value larger than the second detection threshold value, erroneous detection of noise or the like can be more suitably suppressed.

The second detection threshold may be set to a different numerical value depending on the shape of the long defect (for example, the numerical value of Δx / Δy). That is, a plurality of different second threshold values may be set, a threshold value corresponding to the shape of the long defect may be selected, and defective pixels may be redetected from the long defect. In this regard, the third detection threshold is the same.
When the second detection threshold and the first detection threshold are set to the same threshold, the defective pixel is not re-detected with respect to the long defect, and the length detected in the first defect detection using the first detection threshold. A scale defect (provisional defective pixel) may be determined as a defective pixel.

The defect detection means 62 then, for each defective pixel detected with the second detection threshold and the third detection threshold, information indicating the position of the defective pixel on the FPD 56 (the position of the defective pixel on the pixel array of the FPD 56), that is, Create a defective pixel map.
At this time, if the defective pixels detected with the third detection threshold are consecutive in the x direction and / or y direction, the number of pixels exceeding the preset specification limit of the FPD 56, audio output is performed. Or a warning on the monitor 32.

Here, all the defective pixels on the defective pixel map are basically subjected to defective pixel correction in the defect correction unit 64.
However, in the breast cancer inspection apparatus 10 of the present invention, for the defective pixels detected with the second detection threshold, the number of pixels that are consecutive in the x direction and the y direction is detected, and the defective pixel having the larger x direction and y direction is detected. When the continuous number (defect size C) is equal to or larger than a preset third selection threshold (3rd selection Th), information indicating that the defective pixel is a defective pixel that is not subjected to defective pixel correction is displayed as defective pixel. Attach to the map.
That is, if the defect detected with the second detection threshold is “defect size C ≧ 3rd selection Th”, the defect is not subjected to defect pixel correction.
Note that the selection process using the third selection threshold for the defective pixel detected by the second detection threshold may be performed before the creation of the defective pixel map or in parallel with the creation of the defective pixel map. Also good.

In a radiographic image obtained by the breast cancer inspection apparatus 10 or the like, a lesion such as calcification due to breast cancer or the like is usually a point or a substantially circular shape. On the other hand, in the radiographic image, the pattern that is long in the x direction and the y direction is highly likely to be an artifact rather than calcification. That is, even if a pattern caused by a long defect (defective pixels continuous in the x and y directions) is reproduced in the radiographic image, there is a possibility that the doctor misdiagnoses this as a lesion such as calcification. Very low.
On the other hand, the defective pixel correction of the FPD 56 is normally performed by filling holes or the like using adjacent pixels. That is, an image (information) photographed on the defective pixel is removed from the radiation image. Therefore, when a large defective pixel is corrected, on the contrary, the image becomes unnatural, which may cause a misdiagnosis, and when there is information such as calcification in the large defective pixel area. This information cannot be interpreted, and in this respect, there is a possibility of causing a misdiagnosis.

  On the other hand, according to the present invention having the above-described configuration, a large long defective pixel that is extremely unlikely to be misdiagnosed as a lesion such as calcification is reproduced as a radiation image without correcting the defective pixel as it is. By doing so, it is possible to prevent misdiagnosis caused by defective pixels and to significantly suppress misdiagnosis caused by correction of large defective pixels, and to perform more accurate and appropriate breast cancer inspection (inspection by radiographic image). Become.

The third selection threshold value is not particularly limited and may be appropriately determined according to the characteristics of the FPD 56, the accuracy required for the breast cancer inspection apparatus 10, and the like. Alternatively, it may be determined by performing experiments, simulations, or the like, or may be determined by taking into consideration the characteristics of the FPD 56 and the like in the experiments.
Further, the third selection threshold may be set (updated) as appropriate according to the time, the number of shots, the installation environment of the breast cancer inspection apparatus 10, and the like.

Here, the third selection threshold value is preferably a numerical value corresponding to the specification limit of the continuous number of pixels in a long defect, and when a long defect exceeding this is generated, audio output or It is more preferable to output a warning according to the display on the monitor 32. In addition, it is preferable that the specification limit of the long defect is larger than the specification limit of the number of consecutive defective pixels of the normal FPD 56 detected by the third detection threshold.
In other words, the breast cancer inspection apparatus 10 has a small specification limit (so-called standard (default) specification corresponding to the FPD 56) set as the specification limit of the number of defective pixels in the FPD 56 corresponding to the performance and characteristics of the FPD 56. Limit) and a large specification limit corresponding to the long defective pixel are set. For the long defective pixel corresponding to the third selection threshold, a large specification limit is selected and other defective pixels are selected. Select a smaller specification limit.

As described above, a long defective pixel equal to or greater than the third selection threshold is extremely unlikely to cause a misdiagnosis, and in the present invention, it is reproduced as a radiographic image.
Therefore, by having the above configuration, for long defective pixels, it is possible to increase the specification limit until a long defective pixel is generated that would interfere with diagnosis if correction is performed. The life and maintenance frequency of the FPD 56 can be reduced.

When the defect detection unit 62 creates the defective pixel map in this way, the defect detection unit 62 supplies the defective pixel map to the defect correction unit 64.
The defect correction means 64 performs defect pixel correction on the radiation image of the breast M captured by the FPD 56 and processed by the data processing unit 60 with reference to the defective pixel map, and supplies the corrected image to the image processing means 68. Here, the defect correction means 64 basically performs defect pixel correction for all defective pixels on the defective pixel map, but exceeds the third selection threshold and adds information that does not require correction to the defective pixel map. The defective pixel correction is not performed on the defective pixel. That is, the defective pixel is reproduced as a visible image as it is.

  The defective pixel correction method by the defect correction means 64 is not particularly limited, and includes radiation filling such as hole filling by interpolation using adjacent normal pixels and neighboring normal pixels, correction methods described in Patent Document 1 and Patent Document 2, and the like. All known methods for correcting defective pixels of an FPD (radiation image detector) performed in an image capturing apparatus can be used.

Hereinafter, the operation of the breast cancer inspection apparatus 10 will be described.
When a compression plate 48 corresponding to the size of the breast M is mounted and an instruction is given by an engineer, the lifting means 50 moves down the compression plate 48 and compresses the subject's right breast. When the compression of the right breast by the compression plate 48 reaches a predetermined state, radiation is irradiated from the radiation source of the radiation irradiation unit 14, pre-irradiation is performed, and imaging conditions are set. Next, a radiographic image of the breast M is captured according to the imaging conditions, and a radiographic image of the breast M is captured on the FPD 56.

An output signal from the FPD 56 is supplied to the data processing means 60 of the image processing unit 30, where predetermined processing such as log conversion is performed to obtain a radiation image, which is supplied to the defect correction unit 64.
As described above, the defect detection unit 62 acquires a detection image at a predetermined timing, creates (updates) a defective pixel map indicating the position of the defective pixel of the FPD 56, and supplies the defect pixel map to the defect correction unit 64. Yes. The defect correction unit 64 refers to the defective pixel map and corrects the defective pixel described in the defective pixel map. Here, since the defect pixel map includes information indicating that correction is not necessary for defective pixels having pixels that are equal to or greater than the third selection threshold, the defect correction unit 64 stores information indicating that correction is not necessary. No defective pixel correction is performed on the attached defective pixels.

The defect correction unit 64 supplies the radiographic image subjected to the defective pixel correction to the image processing unit 68. The image processing means 68 performs predetermined image processing such as gradation correction and density correction on the supplied radiation image of the breast M, and produces a radiation image (image data) corresponding to the image output by the monitor 32 or the printer 34. , Output to the corresponding part.
The monitor 32 and printer 34 that have received the radiation image reproduce / output the radiation image as a visible image. Here, in the radiographic image of the breast M, defective pixel correction is not performed for defective pixels that exceed the third selection threshold and whose correction pixel information is added to the defective pixel map. Therefore, the defective pixel is reproduced as it is in the reproduced visible image.

  As described above, the defective pixel detection method and the radiographic imaging apparatus of the present invention have been described in detail. However, the present invention is not limited to the above-described embodiments, and various improvements and modifications can be made without departing from the gist of the present invention. Of course, you can do it.

For example, in a breast cancer inspection apparatus, there is no breast in a region away from the chest wall. Therefore, the defective pixel correction may not be performed regardless of the state of the defective pixel in an area that is sufficiently away from the chest wall and where there is no possibility of imaging the breast.
In addition, the above example is an example in which the present invention is used as a suitable example for a breast cancer inspection apparatus. However, the present invention is not limited to this, and a known X-ray such as an X-ray inspection apparatus for photographing a chest is used. All of them can be used in an image capturing device.

  The present invention can be suitably used for defective pixel detection and defect pixel correction of a radiographic image detector in a radiographic imaging apparatus such as a breast cancer inspection apparatus.

DESCRIPTION OF SYMBOLS 10 Breast cancer inspection apparatus 12 Imaging stand 14 Radiation irradiation part 16 Compression means 18 Arm 20 Base 22 High voltage power supply 24 Axis 26, 28 Operation means 30 Image processing part 32 Monitor 34 Printer 48 Compression plate 50 Lifting means 54 Scatter removal grid 56 FPD
60 Data processing means 62 Defect detection means 64 Defect correction means 68 Image processing means 70 Control means

Claims (13)

When detecting a defective pixel of a radiation image detector in which pixels are two-dimensionally arranged in the xy direction,
Obtaining a detection image for detecting defective pixels of the radiation image detector;
Using the detection image to detect a provisional defective pixel with a first detection threshold;
“Δx / Δy” which is a ratio of the number of pixels Δx continuous in the x direction and the number of pixels Δy continuous in the y direction with respect to the temporary defective pixels continuous in at least one of the x direction and the y direction. Calculating and detecting a long defect whose “Δx / Δy” is equal to or lower than the first selection threshold or equal to or higher than the second selection threshold;
For the long defect, a second detection threshold value equal to or greater than the first detection threshold value, and for a temporary defective pixel other than the long defect, a third detection threshold value that is greater than the first detection threshold value. Redetecting defective pixels of the radiation image detector using the detected image;
A step of generating information indicating the position of the defective pixel on the radiation image detector; and detecting the number of pixels consecutive in the x direction and the y direction with respect to the defective pixel detected by the second detection threshold; A step of attaching information indicating that a defective pixel does not require correction to information indicating the position of the defective pixel with respect to a defective pixel having a larger number of consecutive pixels in the direction and y direction than a third selection threshold; A defective pixel detection method characterized by:   The defective pixel detection method according to claim 1, wherein the first selection threshold is 0.8 and the second selection threshold is 1.5.   Two specification limits corresponding to the number of consecutive defective pixels are set. A high specification limit is set for the defective pixels detected by the second detection threshold, and a limit for the defective pixels detected by the third detection threshold is set. The defective pixel detection method according to claim 1, wherein a low specification limit is selected.   The defective pixel detection method according to claim 3, wherein a specification limit corresponding to a defective pixel detected by the third detection threshold is a specification limit set as a standard for the radiation image detector.   The defective pixel detection method according to claim 3 or 4, wherein the third selection threshold corresponds to a specification limit of a defective pixel detected by the second detection threshold.   The defective pixel detection method according to claim 3, wherein a warning is output when the number of consecutive defective pixels exceeds a specification limit.   Furthermore, the defective pixel detection method in any one of Claims 1-6 which correct | amend defective pixels other than the defective pixel to which the information of the said defective pixel which does not need correction | amendment was attached | subjected. A radiation source;
a radiation image detector in which pixels are two-dimensionally arranged in the xy direction;
A detection image for detecting a defective pixel of the radiation image detector is acquired, and a temporary defective pixel is detected by a first detection threshold using the detection image, and at least one of the x direction and the y direction is detected. Is calculated by calculating “Δx / Δy”, which is the ratio of the number of pixels Δx continuous in the x direction and the number of pixels Δy continuous in the y direction. ”Is detected as a long defect that is less than or equal to the first selection threshold value or greater than or equal to the second selection threshold value. For the provisional defective pixel, a defective pixel of the radiological image detector is re-detected using the detection image with a third detection threshold value that is larger than the first detection threshold value, and the radiographic image detector has the Defect position information indicating the position of the defective pixel And detecting the number of consecutive pixels in the x direction and the y direction with respect to the defective pixel detected by the second detection threshold, and selecting the number of consecutive pixels in the larger x direction and y direction as the third number. A defective pixel detection unit that adds information indicating that the defective pixel is not necessary to be corrected to the defective position information for a defective pixel that is equal to or greater than a threshold;
In accordance with the defect position information created by the defective pixel detection means, the defective pixels other than the information indicating that they are defective pixels that do not require correction are corrected for the radiographic image captured by the radiographic image detector. A radiographic imaging device comprising defective pixel correction means.   The radiographic image capturing apparatus according to claim 8, wherein the first selection threshold is 0.8 and the second selection threshold is 1.5.   Two specification limits corresponding to the number of consecutive defective pixels are set, a high specification limit is detected for the defective pixels detected by the second detection threshold, and a low specification limit is set for the defective pixels detected by the third detection threshold. The radiographic imaging device according to claim 8 or 9, wherein the specification limit is selected.   The radiographic image capturing apparatus according to claim 10, wherein a specification limit corresponding to a defective pixel detected by the third detection threshold is a specification limit set as a standard for the radiographic image detector.   The radiographic imaging apparatus according to claim 10 or 11, wherein the third selection threshold corresponds to a specification limit of a defective pixel detected by the second detection threshold.   The radiographic imaging apparatus according to claim 10, wherein a warning is output when the number of consecutive defective pixels exceeds a specification limit.

Images