JP2004194729A - Supporting device for image-based diagnosis - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a supporting device for an imaged-based diagnosis with which a shadow of a possible lesion is detected quickly and appropriately by inputting only a medical image which is suitable for detecting the shadow of the directing possible lesion. <P>SOLUTION: The supporting device for the image-based diagnosis is constituted to judge whether the inputted medical image is suitable for detecting the shadow of the directing possible lesion or not based on at least either a related information of the inputted medical image or an analysis result of the inputted medical image and perform the detection quickly and appropriately by inputting only the medical image suitable for detecting the shadow of the directing possible lesion. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

次のステップ３０６では、ステップ３０４で決定されたしきい値で２値化処理を行い、その後ステップ３０８へ進んでラベリング処理を行う。
【００４９】
ステップ３１０では、ステップ３０８でラベリング処理した結果に対して、画素数が最大となる領域（例えば全肺野領域を含む断層画像であれば、全肺野領域が画素数最大となる）のみを抽出し、次のステップ３１２で上記領域に対する穴埋め処理を行う。
【００５０】
ステップ３１４では、ステップ３１０で得られた画素数最大となる領域の画像（Ｔ(x,y)とする）と該画像に対してステップ３１２での穴埋め処理により得られた画像（Ｆ(x,y)とする）とを、以下の式に示すように排他的論理和により合成処理する。
【００５１】
【数２】

上記合成処理により得られた画像Ａ(x,y)は、上述の画素数最大の領域に囲まれた領域のうち、画素値がしきい値以下の領域である。この領域の数および面積は、撮影対象の臓器や部位等により異なるものであり、例えば画像が全肺野領域を含む断層画像である場合、上記領域は気管等を示す。
【００５２】
次のステップ３１６では、上記合成処理により得られた領域の面積があらかじめ定めた範囲内であるか否かを判断し、肯定されるとステップ３１８へ進み、否定されるとステップ３２４へ進んで処理フラグを”面積：設定範囲外”とした後、本サブルーチンを終了してリターンする。
【００５３】
ステップ３１８では、ステップ３１４での合成処理により得られた領域の数が適正であるか否かを判断する。肯定されるとステップ３２０へ進んで処理フラグを”解析結果：正”とした後に本サブルーチンを終了してリターンし、否定されるとステップ３２２へ進んで処理フラグを”領域数：不正”とした後に本サブルーチンを終了してリターンする。
【００５４】
なお、上記ステップ３１６および３１８において判断の対象となる、合成処理により得られた領域の面積および数の適正な値は、医用画像の撮影対象となる肺や脳等の臓器の種類および撮影部位等を考慮することにより設定することができる。
【００５５】
次に、上述のステップ１２６の詳細を説明する。図２のステップ１２６におけるメッセージ表示処理が実行されると、図５に示す処理サブルーチンが実行され、ステップ４００へ進む。
【００５６】
ステップ４００では図２の処理フローの結果設定された処理フラグを読み取り、次のステップ４０２で該処理フラグに応じたメッセージをＣＲＴ４８に表示する。表示されるメッセージは、図６の(ａ)、(ｂ)の例に示すように、「撮影機種が違います」、「撮影部位が違います」等がある。
【００５７】
以上説明したように、本実施の形態が適用された画像診断支援システム１０では、画像診断支援装置２０に入力された医用画像が目的とする病巣候補陰影の検出に適合しているか否かを、該入力された医用画像のヘッダ情報と該入力された医用画像の画像解析結果とに基づいて判定するので、目的とする病巣候補陰影の検出に適合した医用画像のみを検出対象として入力し、医用画像中に病巣候補陰影を識別可能に表示する等、特許文献１に記載された全ての処理を迅速かつ適切に行うことができる。
【００５８】
さらに、病巣候補陰影検出の結果が正常であるか否かを判定し、異常な検出結果を排除するので、正常な検出結果のみを登録し、表示することができる。
【００５９】
【発明の効果】
以上説明したように、本発明によれば、目的とする病巣候補陰影の検出に適合する医用画像のみを検出対象として入力し、病巣候補陰影の検出を迅速かつ適切に行うことができる、という効果がある。
【図面の簡単な説明】
【図１】本発明の一の実施の形態に係る画像診断支援システムの全体構成図である。
【図２】本発明の一の実施の形態に係り、画像診断支援装置での処理の概要を示すフローチャートである。
【図３】本発明の一の実施の形態に係り、ヘッダ情報抽出処理を示すフローチャートである。
【図４】本発明の一の実施の形態に係り、画像解析処理を示すフローチャートである。
【図５】本発明の一の実施の形態に係り、メッセージ表示処理を示すフローチャートである。
【図６】本発明の一の実施の形態に係り、メッセージ表示の例を示すイメージ図である。
【符号の説明】
１０・・・画像診断支援システム、２０・・・画像診断支援装置、５８・・・ネットワーク、６０・・・画像サーバ[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an image diagnosis support apparatus that determines the suitability of a medical image for detecting a lesion candidate shadow.
[0002]
[Prior art]
The detection of a lesion candidate shadow is performed by inputting a medical image (an image captured by a medical image diagnostic apparatus such as a CT apparatus or an MRI apparatus) to a lesion candidate shadow detection apparatus, as described in Patent Document 1 below. At this time, in the related art, the operator of the lesion candidate shadow detection device selects a medical image and inputs the medical image to the lesion candidate detection device.
[0003]
[Patent Document 1]
International Publication No. WO 02/02002
[Problems to be solved by the invention]
However, in the above-described conventional technology, since the medical image input to the lesion candidate shadow detection device is based on the selection of the operator, an image that does not match the lesion candidate shadow detection device due to an operation error is selected and input. As a result, the detection of a lesion candidate shadow may not be performed properly. For example, when performing image processing for automatically detecting a lesion candidate shadow from a tomographic image (CT image) including the entire lung field, if a tomographic image of a type different from the inspection target is selected and input, an appropriate There is a problem in that image processing cannot be performed and useless processing time is consumed.
[0005]
The present invention has been made in view of the above circumstances, and inputs only a medical image that matches detection of a target lesion candidate shadow as a detection target, and performs image diagnosis support in which detection of a lesion candidate shadow is performed quickly and appropriately. It is intended to provide a device.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, an image diagnosis support apparatus according to the present invention according to claim 1, wherein input means for inputting a medical image, and the medical image input by the input means is adapted for detection of a lesion candidate shadow Determining means for determining whether or not the input medical image is based on at least one of the information related to the input medical image and the analysis result of the input medical image; A lesion candidate shadow detecting means for detecting a lesion candidate shadow for a medical image based on detection of a lesion candidate shadow, and a display for displaying the lesion candidate shadow detected by the lesion candidate shadow means in the medical image in a distinguishable manner. Means.
[0007]
According to the first aspect of the present invention, the medical image is input by the input unit. The medical image is an image captured by a medical image diagnostic apparatus such as a CT apparatus, an MRI apparatus, and an ultrasonic diagnostic apparatus. Whether the input medical image is a medical image suitable for detection of a lesion candidate shadow based on at least one of information related to the medical image and an analysis result of the input medical image by the determination unit It is determined whether or not.
[0008]
The information related to the medical image includes the image file name of the medical image, the presence / absence of a header portion of the image file, the imaging device and the imaging time included in the header portion, the patient name, and the like. The determination of the medical image can be made based on whether or not the medical image matches.
[0009]
Further, in the analysis of the medical image, the region where the image is the target of the lesion candidate shadow detection is determined based on whether or not the area of the region having the maximum number of pixels in the input medical image is within a preset range. There is a method of determining whether or not the image includes the image.
[0010]
By using the method as described above, the determination unit determines whether or not the input medical image is a medical image that is suitable for detecting the target lesion candidate shadow. Can be excluded from the processing target, and the focus candidate shadow detection means detects a focus candidate shadow only for an appropriate medical image. Then, the detected lesion candidate shadow is identifiably displayed in the medical image. This means is described in detail in Japanese Patent Application Laid-Open Publication No. HEI 10-163, and the description thereof is omitted here.
[0011]
In this manner, whether or not the input medical image is suitable for the detection of the target lesion candidate shadow is determined based on the information related to the input medical image and the analysis result of the input medical image. Since the determination is made based on at least one, only medical images that match the detection of the target lesion candidate shadow can be input as detection targets, and processing can be performed quickly and appropriately.
[0012]
The determining means may further include a detection result determining means for determining whether a detection result by the lesion candidate shadow detecting means is normal.
[0013]
The detection result of the lesion candidate shadow includes the number and size of the lesions, and these factors have a certain tendency depending on the type and site of the organ from which the image was captured, the patient to be captured, and the like.
[0014]
Therefore, the determination as to whether or not the focus detection result is normal can be made by comparing the detection result of the focus candidate shadow with a predetermined reference in consideration of the above tendency.
[0015]
As described above, by judging whether or not the detection result is normal, an abnormal detection result of a lesion candidate shadow due to an incorrect image or the like can be excluded, and only a correct detection result can be obtained.
[0016]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, preferred embodiments of an image diagnosis support apparatus according to the present invention will be described in detail with reference to the accompanying drawings.
[0017]
FIG. 1 shows an overall configuration of an image diagnosis support system 10 to which the present embodiment is applied. The image diagnosis support system 10 includes an image diagnosis support device 20, an image server 60, and a network 58.
[0018]
The image diagnosis support apparatus 20 is an embodiment of the image diagnosis support apparatus according to the present invention, and the image server 60 stores and manages medical images. The image diagnosis support device 20 and the image server 60 transmit and receive image data via the network 58.
[0019]
Note that the image diagnosis support apparatus 20 can process medical images captured under various conditions regarding the types and parts of organs such as lungs and brain. Here, the medical image is an image captured by a medical image diagnostic apparatus such as a CT apparatus, an MRI apparatus, and an ultrasonic diagnostic apparatus.
[0020]
Various networks such as a LAN (local area network) and the Internet can be used as the network 58. One or more servers connected to the network can be used as the image server 60.
[0021]
The image diagnosis support apparatus 20 includes a central processing unit (hereinafter, referred to as a CPU) 40 that controls each component, and the CPU 40 is connected to a common bus 56. The common bus 56 stores a control program of the image diagnosis support apparatus 20, and is used as an area for temporarily storing data for image processing, and a medical image file, a database, message data, A magnetic disk 44 storing programs and the like, a display memory 46 for temporarily storing image data for display, a controller 52, a keyboard 54 having keys and switches for setting various parameters, a speaker 57, A communication interface device 55 that communicates with the image server 60 via the network 58 is connected. The display memory 46 is connected with a display (hereinafter referred to as CRT) 48 for displaying an image, a message corresponding to the processing result, and the like. The controller 52 is a means for performing various operations on the screen of the CRT 48. Mouse 50 is connected.
[0022]
As a storage device, a device such as a CD-ROM drive, a magneto-optical disk (MO) drive, or a DVD drive can be connected in addition to the magnetic disk 44, and a storage medium that can be inserted into and removed from these devices. The input and output of an image file and the like may be performed via the.
[0023]
Next, the operation of the above embodiment will be described. First, an outline of a processing routine executed by the image diagnosis support apparatus 20 will be described with reference to FIG.
[0024]
First, in step 100, an image file of a medical image corresponding to a condition set in advance by the mouse 50 or the keyboard 54 is read from the magnetic disk 44.
[0025]
The image file may be read from the magnetic disk 44, read from the image server 60 via the network 58, or provided with the storage device described above in the image diagnosis support apparatus 20, and read via the storage medium. Good.
[0026]
In the next step 102, header information is extracted for one file to be processed in the following steps among the read files (described later), and the process proceeds to step 104. Here, the header information is information related to the image included in the image file. For example, in the case of a DICOM format image file, the header information includes a diagnostic device and a photographing time.
[0027]
In step 104, it is determined whether the header information is correct based on the result of the header information extraction in step 102. If affirmed, the process proceeds to step 106 to perform image analysis. If denied, the process proceeds to step 126 to display a message (to be described later), and the processing for one file is completed.
[0028]
By the processing of steps 102 and 104, based on the header information, image files that do not match the detection of the target lesion candidate shadow can be excluded from detection targets.
[0029]
If the determination in step 104 is affirmative, image analysis is performed in step 106 (described later), and in the next step 108, it is determined whether the analysis result is normal. If affirmed, the process proceeds to step 110 to detect abnormal shadows and lesions, and if denied, the process proceeds to step 126 to display a message (described later) and terminate the processing for one file.
[0030]
By the processing of steps 106 and 108, based on the result of the image analysis, an image file that does not match the detection of the target lesion candidate shadow can be excluded from detection targets.
[0031]
As described above, by performing the processing of steps 102 to 108, it is possible to exclude an inappropriate image file as a target of detection of a target lesion candidate shadow and to input only an appropriate image file as the input for detection of a lesion candidate. In addition, this makes it possible to reduce the processing time as a whole.
[0032]
In step 110, an abnormal shadow and lesion detection process is performed. This processing is to determine the type of the site or organ in order to perform an optimal detection process corresponding to the site or the type of the organ to be diagnosed, and perform various image processing on the result of the multi-level processing, A lesion candidate shadow (a shadow considered to be a lesion), that is, an abnormal shadow is detected.
[0033]
For the details, a method described in WO 02/02002 in which automatic extraction and judgment / deletion are performed for each property of a shadow to be extracted and for each existing location is adopted. By performing the focus candidate shadow detection processing based on the multi-valued image as in this method, it is possible to reduce the time required for computation by a computer and the like, and reduce the processing load. Then, the shadow determined to be the focus as a result of the processing is left, and the shadow that is not determined is deleted.
[0034]
At the time of the detection process, as shown in the example of FIG. 6C, the process of processing the image may be displayed on the CRT 48.
[0035]
In the next step 112, the number of lesions detected as a result of the processing in step 110 is extracted, and the routine proceeds to step 114, where it is determined whether or not the number of detected lesions is within a preset range. If affirmed, the process proceeds to step 116 to register the detection result in the database, and if denied, the process proceeds to step 124 to set the processing flag to "detection number: out of the setting range", and proceeds to step 126 to respond to the processing flag. Displays the message that was sent.
[0036]
Here, the processing flag is a flag in which different contents are set according to each processing result, and the message display in step 126 is performed according to the processing flag (described later).
[0037]
Note that, as the number of detections used as a criterion for determination in step 114, a value in a range that can be regarded as normal can be set, for example, based on the number of past detections of the patient to be detected.
[0038]
As described above, by the processing of steps 112 and 114, an abnormal detection result due to an incorrect image file, for example, a detection result due to an image file of another patient or the like can be excluded, and only a correct detection result can be obtained.
[0039]
In step 118, the detection result is displayed on the CRT 48, and thereafter, the process proceeds to step 120, and it is determined whether or not the processing has been completed for all the image files read in step 100. If affirmed, the process proceeds to step 122; if denied, the process returns to step 102 to start processing the next image file.
[0040]
In step 122, it is determined whether or not to end the processing. This determination can be made based on the presence or absence of an input indicating the end of the processing by the keyboard 54 or the mouse 50. If the determination is affirmative, the present processing routine is ended. If the determination is negative, the process returns to step 100.
[0041]
Next, the details of step 102 will be described. When the header information extraction processing in step 102 in FIG. 2 is executed, a processing subroutine shown in FIG. 3 is executed, and the process proceeds to step 200.
[0042]
In step 200, the type of the diagnostic device is read from the header of the image file, and the flow advances to step 202 to determine whether or not the type of the read diagnostic device matches a preset one. If the determination is affirmative, the process proceeds to step 204, where the imaging part is read from the header of the image file.
[0043]
In step 206, it is determined whether or not the imaging site read in step 204 is the target site of the target lesion candidate detection. If affirmative, the process proceeds to step 208, where the photographing time is further read from the header of the image file.
[0044]
In step 210, it is determined whether or not the photographing time read in step 208 is within a preset time range. If affirmed, the process proceeds to step 212, where the processing flag is set to "header information: correct", and then this subroutine is terminated and the process returns. If not, the process proceeds to step 214 and the processing flag is set to "imaging time: incorrect". After the setting, the subroutine ends and the routine returns.
[0045]
Next, the details of step 106 will be described. When the image analysis processing in step 106 in FIG. 2 is executed, a processing subroutine shown in FIG. 4 is executed, and the process proceeds to step 300.
[0046]
In step 300, a histogram of the pixel values of each pixel is created for the image data in the image file, and then the process proceeds to step 302, where it is determined whether or not the pixel values are continuously distributed based on the histogram. If affirmed, the process proceeds to step 304, and if denied, the process proceeds to step 326 to set the processing flag to "pixel value: discontinuous", and thereafter ends this subroutine and returns.
[0047]
In step 304, a threshold value is determined based on the histogram created in step 300. The threshold value is a criterion value of a pixel value for extracting only an area indicating a subject such as a patient from the entire area of a captured image. The pixel value when the following expression is satisfied is used as the threshold value.
[0048]
(Equation 1)

In the next step 306, binarization processing is performed using the threshold value determined in step 304, and thereafter, the process proceeds to step 308 to perform labeling processing.
[0049]
In step 310, only the region having the maximum number of pixels (for example, if the tomographic image includes the entire lung field region, the entire lung field region has the maximum number of pixels) is extracted from the result of the labeling process in step 308. Then, in the next step 312, a hole filling process for the above-mentioned area is performed.
[0050]
In step 314, an image (T (x, y)) of the region having the maximum number of pixels obtained in step 310 and an image (F (x, y)) are combined by exclusive OR as shown in the following equation.
[0051]
(Equation 2)

The image A (x, y) obtained by the synthesis processing is an area where the pixel value is equal to or less than the threshold value among the area surrounded by the area having the maximum number of pixels. The number and area of the regions differ depending on the organ or site to be photographed. For example, when the image is a tomographic image including the entire lung field region, the region indicates a trachea or the like.
[0052]
In the next step 316, it is determined whether or not the area of the region obtained by the combining process is within a predetermined range. If the result is affirmative, the process proceeds to step 318, and if the result is negative, the process proceeds to step 324. After setting the flag to "area: out of the set range", the subroutine is terminated and the routine returns.
[0053]
In step 318, it is determined whether the number of regions obtained by the combining process in step 314 is appropriate. If affirmed, the process proceeds to step 320, where the processing flag is set to "analysis result: correct", and then this subroutine is terminated and the process returns. If denied, the process proceeds to step 322 and the processing flag is set to "number of areas: invalid". After this subroutine ends, the process returns.
[0054]
The appropriate values of the area and the number of the regions obtained by the synthesis processing, which are to be determined in the above steps 316 and 318, are determined based on the types of organs such as lungs and brains to be photographed for medical images and the photographed parts. Can be set.
[0055]
Next, the details of step 126 will be described. When the message display processing in step 126 in FIG. 2 is executed, a processing subroutine shown in FIG. 5 is executed, and the process proceeds to step 400.
[0056]
In step 400, the processing flag set as a result of the processing flow of FIG. 2 is read, and in the next step 402, a message corresponding to the processing flag is displayed on the CRT 48. The displayed messages include “imaging model is different” and “imaging area is different” as shown in the examples of FIGS. 6 (a) and 6 (b).
[0057]
As described above, in the image diagnosis support system 10 to which the present embodiment is applied, it is determined whether or not the medical image input to the image diagnosis support device 20 is suitable for detecting a target lesion candidate shadow. Since the determination is made based on the header information of the input medical image and the image analysis result of the input medical image, only a medical image suitable for detecting the target lesion candidate shadow is input as a detection target, and All the processes described in Patent Literature 1 can be performed quickly and appropriately, such as displaying lesion candidate shadows in an image in a distinguishable manner.
[0058]
Further, it is determined whether or not the result of the focus candidate shadow detection is normal, and the abnormal detection result is excluded, so that only the normal detection result can be registered and displayed.
[0059]
【The invention's effect】
As described above, according to the present invention, it is possible to input only a medical image suitable for detecting a target lesion candidate shadow as a detection target, and to quickly and appropriately detect a lesion candidate shadow. There is.
[Brief description of the drawings]
FIG. 1 is an overall configuration diagram of an image diagnosis support system according to an embodiment of the present invention.
FIG. 2 is a flowchart showing an outline of processing in the image diagnosis support apparatus according to one embodiment of the present invention.
FIG. 3 is a flowchart illustrating a header information extraction process according to an embodiment of the present invention.
FIG. 4 is a flowchart illustrating an image analysis process according to an embodiment of the present invention.
FIG. 5 is a flowchart showing a message display process according to one embodiment of the present invention.
FIG. 6 is an image diagram showing an example of a message display according to an embodiment of the present invention.
[Explanation of symbols]
10 image diagnostic support system, 20 image diagnostic support device, 58 network, 60 image server

Claims (1)

Input means for inputting a medical image,
Whether or not the medical image input by the input unit is a medical image suitable for detection of a lesion candidate shadow is at least one of information related to the input medical image and an analysis result of the input medical image. Determining means for determining based on one of the
Lesion candidate shadow detection means for detecting a lesion candidate shadow for a medical image adapted to detection of a lesion candidate shadow based on the determination result of the determination means,
Display means for displaying a lesion candidate shadow detected by the lesion candidate shadow detection means in the medical image in an identifiable manner.

Images