JP5543871B2 - Image processing device - Google Patents

Description

  The present invention relates to an image processing apparatus, for example, a method for automatically setting an image processing condition for extracting a lesioned part in an image processing apparatus for a contrast image obtained by photographing a measurement object with a tomography apparatus.

  In recent years, medical images and medical information have been increasing remarkably with the progress of image diagnostic apparatuses and the like, and the amount of accumulation has been increasing. This causes an extreme increase in the burden on clinicians and radiologists who make diagnoses using medical images, and as a result, accumulated medical images and medical information are not utilized. For this reason, various types of image processing have been performed for the purpose of improving the quality of diagnosis and treatment by effectively using medical images.

  The conventional technique is an area for obtaining the reference signal value first, which can perform gradation processing suitable for firstly examining a change in a medical condition between the same patients for the above-described purpose. The image processing apparatus can automatically recognize a thoracic vertebra part without using information on a lung field. That is, the above-described conventional technique uses a signal value of a region that always maintains the same density, for example, a thoracic vertebra region or a shoulder region, regardless of the lesion in the chest radiographic image as a representative value, and other regions based on the representative value This is an image processing apparatus that automatically determines the gradation conversion processing conditions by observing the change in the image and performs the gradation processing (see, for example, Patent Document 1).

  In addition, there is a technique for performing image processing using a thoracic vertebra region as in the above-described conventional technique (see, for example, Patent Document 2).

JP 2000-079110 JP2007-300966

  However, the above-described conventional method cannot be applied to a contrast image in which the luminance value of the image changes depending on the imaging time of the same patient. A contrast image is a general method used to clarify the contrast of a tumor such as cancer, and examines the gradation of a contrast agent administered to a patient near the tumor. That is, the contrast agent is carried to an organ through a blood vessel such as an artery, and if there is a tumor in the organ, it stays in the vicinity. Therefore, there is a problem that the contrast medium stays different depending on the blood flow volume and imaging time of the patient and the gradation of the contrast image changes, and the lesion part is automatically recognized and automatically extracted by the gradation change. It becomes difficult.

  Accordingly, an object of the present invention has been made in view of such a problem, and firstly, it is to provide an image processing apparatus capable of automatically extracting a lesion, and secondly, Another object of the present invention is to provide an image processing apparatus capable of automatically setting, from the gradation of an artery, an image processing condition that can appropriately extract a contrast image that changes depending on a blood flow volume of an object and an imaging time.

  In order to solve the above problems, an image processing apparatus according to the present invention has the following configuration.

  (1) An input unit that captures a medical image, an image processing unit that extracts a lesion from the medical image, and a display unit that displays an image of the lesion extracted by the image processing unit. Is a contrast image obtained by administering a contrast agent to a subject and photographing a measurement object with a tomography apparatus, and the brightness of another tissue in the contrast image that is different from a lesion that changes in brightness due to the contrast agent And a means for automatically extracting a lesion from a contrast image based on the set automatic extraction parameter based on the set automatic extraction parameter. To do.

  (2) In the image processing apparatus of (1), the other tissue is an artery.

  (3) In the image processing apparatus according to (2), the processing means inputs a contrast image and extracts a feature amount from the gradation of the artery in the image, a feature amount and an automatic extraction parameter. It includes a parameter selector that selects a parameter suitable for automatic extraction of a lesion using reference data serving as a reference for setting, and an image extractor that extracts a lesion based on the parameter.

  (4) In the image processing apparatus of (2), the gradation of the artery is an average value of a histogram of the gradation of the artery.

  (5) The image processing apparatus according to (3) is characterized in that the automatic extraction of a lesioned part by an image extractor is executed by a level set algorithm, and the automatic extraction parameter is a gradation of a lesioned part.

  (6) The image processing apparatus of (3) is characterized by having means for presenting reference data to a diagnostician and input means for changing the reference data by the diagnostician.

  (7) In the image processing apparatus of (1) or (2), the reference data for setting the automatic extraction parameter is updated based on the feature amount output by the arterial feature extractor and the parameter output by the parameter selector. It has the means.

  According to the above configuration, the contrasted image changes with the lesioned part, and the lesioned part is automatically detected from the contrasted image using a feature other than the lesioned part other than the lesioned part, for example, the feature amount of the artery. To extract. That is, the monitoring position can be determined with high accuracy, and an appropriate lesion can be extracted based on the feature amount that changes in conjunction with the current state of the subject.

  According to the present invention, when used for diagnosis of a lesioned part of a subject, the lesioned part is automatically extracted without inputting information necessary for the diagnostician to automatically extract the lesioned part to the image processing apparatus. can do.

1 is a block diagram showing a first embodiment of an image processing apparatus of the present invention. It is a correlation diagram of an arterial feature and parameters for automatic extraction. It is a flowchart figure which shows 2nd embodiment of the image processing apparatus of this invention. It is a block diagram which shows 3rd embodiment of the image processing apparatus of this invention. It is a block diagram which shows 4th embodiment of the image processing apparatus of this invention. It is a flowchart figure which shows 4th embodiment of the image processing apparatus of this invention. It is a figure which shows an example of the screen shown to the diagnostician shown in 4th embodiment. It is a figure explaining the figure shown in FIG.

Embodiments will be described below with reference to the drawings.
<First embodiment>
FIG. 1 is a block diagram showing a first embodiment of the image processing apparatus of the present invention. In FIG. 1, medical images such as contrast images are stored in the database server 101. In this embodiment, contrast CT (abbreviation of computer tomography, hereinafter referred to as CT) image 110_0 and contrast CT image 110_1 are stored. Think if you are.

  When a diagnostician executes image processing on the terminal PC 102_0, an execution command is transmitted to the image processing apparatus 100 via the network 204. The image processing apparatus 100 acquires the corresponding contrast CT image (CT image) 110_0 or 110_1 from the database server 101 via the network 200 in accordance with the transmitted execution command. The image processing apparatus 100 automatically extracts a lesion from the acquired contrast CT image. (1) Artery feature extractor 120, (2) Parameter selector 121, (3) Image The following processing (1) to (5) is performed in each processing means of the filter device (Image filter) 122, (4) Image extractor (Segmentation) 123, and (5) Post image processor (Post image processor) 124. . In the following description, the symbols in FIGS. 1 and 2 are used together.

  (1) The arterial feature quantity extractor 120 receives, for example, the contrast CT image 110_0 and extracts the feature quantity from the gradation of the artery in the image. Here, the feature amount indicates the intensity of brightness obtained from the artery. Since the arterial gradation has a plurality of gradations instead of a single gradation, the arterial feature quantity extractor 120 obtains a histogram of the arterial gradation and outputs the average value as the arterial feature quantity.

  (2) The parameter selector 121 receives the arterial feature amount output from the arterial feature amount extractor 120 and the reference data (Criterion data) 111, and selects a parameter suitable for automatic extraction of a lesion. FIG. 2 shows the correlation between the arterial feature quantity, the reference data 111, and the parameter (luminance value of the lesion) output from the parameter selector 121.

In FIG. 2, the horizontal axis indicates the arterial feature amount, the vertical axis indicates the luminance value of the lesion part that is the output parameter, the solid line 300 in the figure indicates the reference data, and the data points 301 are obtained from a plurality of contrast images. The optimum parameters obtained experimentally are shown. The parameter selector 121 receives the arterial feature amount and outputs a parameter (luminance value of the lesion) that intersects the reference data 300. That is, when the reference data is set as the function f _{reference data} , it can be expressed as shown in Expression (1).

Automatic extraction parameter = f _{reference data} (arterial feature) (1)
Further, the correlation is obtained with the optimum parameter experimentally obtained from FIG. 2, and the parameter can be automatically set by using the present invention.
The correlation between the feature amount (luminance value) of the artery and the lesioned part is indicated by a gradation correlation line 302.

  This reference data can be updated by a user (such as a diagnostician) of the apparatus, and examples of the update function will be described in the third and fourth embodiments.

  (3) The image filter 122 executes a general image filter, for example, gradation processing.

(4) The image extractor 123 extracts a lesion part based on the parameter. As described above, since the parameter is set in accordance with the gradation of the contrast CT image 110_0, the diagnostician does not need to set a parameter necessary for extracting the lesioned part and can automatically extract the parameter. Further, the image extractor 120 has a function of enhancing and outputting the extracted lesioned part to the contrast CT image 110_0 so that the diagnostician can understand.
Note that the contrast CT image has been described by taking 110_0 as an example, but the above processing can also be applied to 110_1.

  (5) In the post-process 124, a process using a general image filter, for example, a morphology filter process for fine adjustment of the extraction result is executed.

  The image processing apparatus 100 performs the processes (1) to (5) described above, and transfers the highlighted contrast CT image via the network 203 to the terminal PC 102_0 where the diagnostician has performed the image processing. The contrast CT image is displayed on the display means 103_0 provided in the terminal PC102_0, and the image is presented to the diagnostician. The above description has been given for the terminal PC 102_0 and the display unit 103_0. Needless to say, the same applies to the terminal PC 102_1 and the display unit 103_1.

  As a result, it is possible to realize an image processing apparatus having a function of automatically extracting a lesion without requiring input by a diagnostician. In general, it is difficult to visually recognize the luminance change of the artery on the screen, but by automating as in the present embodiment, it is possible to accurately grasp a plurality of areas and luminance values.

In the present embodiment, the arterial feature amount may be any value that can be used for determining the blood flow rate and imaging time of the subject, and the difference between the gradation of the artery and the gradation of the organ to be extracted may be used as the feature amount.
In addition, the brightness of other tissues whose brightness varies depending on the contrast agent may be used, not necessarily the artery.
<Second Embodiment>
Next, a second embodiment of the present invention will be described.
In the second embodiment, an embodiment in which the image extraction algorithm in the first embodiment is a level set algorithm 450 is shown. FIG. 3 shows a flowchart of lesion portion extraction in the image processing apparatus when the level set algorithm 450 is used as an extraction algorithm. The level set algorithm 450 is an algorithm generally used in image segmentation.

  In FIG. 3, after starting lesion extraction 400, preprocessing 401 is processed by the arterial feature quantity extractor 120 and the parameter selector 121, the image filter 402 is processed by the image filter 122, and auxiliary function initialization 403. The update amount threshold determination 406 is processed by the image extractor 123, and the post-processing 407 is processed by the post-processor 124.

  The artery feature quantity is extracted by the arterial feature quantity extractor 120, and the parameter selector 121 sets the parameters of the velocity function calculation 404 according to the reference data shown in FIG. Next, gradation processing is performed in the image filter 402. Thereafter, an auxiliary function initialization 403 which is a general process is processed by the level set algorithm. Next, in the speed function calculation, a speed function is calculated according to the parameters set in the preprocessing 401. The speed function is calculated from the gradation value of the contrast CT image, the differential value of the gradation, and the shape of the gradation, and the parameter is used as a reference value for the gradation value. Since a predetermined range of gradations is extracted as a lesioned part around the reference value, it is appropriate to set the reference value as an extraction target, here a representative value of the gradation value of the affected part. The representative value changes depending on the blood flow volume of the subject and the imaging time. However, since the reference value is appropriately set by the parameter selector, it is possible to automatically respond to the change. After that, the auxiliary function is updated by the auxiliary function update 405 in FIG. 3 according to the speed function, and when the update amount is compared with a predetermined threshold value in 406 and the update amount falls below the threshold value, post-processing such as image processing is performed. After 407, the extraction ends.

In this embodiment, the image extraction algorithm may be any algorithm that can automatically set extraction parameters, and may be snakes or region growing using the gradation value to be extracted. Further, 402 and 407 may be any one that performs appropriate image processing, and may be, for example, noise removal processing, smoothing processing, edge enhancement processing, morphological filter, or the like. Further, the value selected by the parameter selector may be an image processing condition for performing appropriate extraction, and may be, for example, a threshold value of a differential value of gradation other than the reference value.
<Third embodiment>
Next, a third embodiment of the present invention will be described.
In the third embodiment, the image processing apparatus of the first embodiment further includes a reference data updater 125 that updates the reference data 111. FIG. 4 is a block diagram of the third embodiment.
In addition, the same code | symbol is attached | subjected to the part equivalent to 1st embodiment.

  In the first embodiment, the reference data 111 is set in advance. On the other hand, in the third embodiment, the reference data is updated by a user (such as a diagnostician) of the apparatus. An image processing apparatus.

  In FIG. 4, a reference data updater 125 having a function of updating the reference data 111 replaces new reference data 111_1 transmitted from the terminal PC 102_0 via the network 204 with the old reference data 111. As a result, in addition to the features of the first embodiment, even when a medical image such as a CT device is handled by different imaging devices, the reference data can be used as a new imaging device even when the conventional reference data cannot be appropriately extracted. By replacing with reference data corresponding to, appropriate extraction can be maintained.

In the present embodiment, as described above, the terminal PC102_0 is used as a means for inputting the new reference data 111_1. However, any device that replaces the reference data 111 so far with the new reference data 111_1 may be used. The data may be directly input to 100 or stored in the database server 101 and transferred from the database server to the image processing apparatus.
<Fourth embodiment>
Next, a fourth embodiment of the present invention will be described with reference to FIGS.
As shown in FIG. 5, in the fourth embodiment, the image processing apparatus of the third embodiment is further provided with an automatic reference data updater (Auto update) 126 that automatically updates the reference data 111.
In addition, the same code | symbol is attached | subjected to the part equivalent to 3rd embodiment.

  As described in the third embodiment, the reference data 111 can be updated from the terminal PC 102_0. On the other hand, the fourth embodiment describes an image processing apparatus that automatically updates the reference data 111 from the extraction result, that is, learns the reference data.

  FIG. 5 shows a block diagram of the fourth embodiment. In addition to the functions described in the third embodiment, the parameter selector 121 in FIG. 5 includes a function 203 that transmits parameters to the terminal PC 102_0 or 102_1, and a function 205 that can reset parameters from the terminal PC. The automatic reference data updater 126 receives the arterial feature quantity from the arterial feature quantity extractor 120 and the parameter for automatic extraction from the parameter selector 121, and updates the reference data 111 based on the inputted parameter. It has a function.

  FIG. 6 shows a flowchart of the fourth embodiment. In addition, the same code | symbol is attached | subjected to the part equivalent to FIG. Also, the image extraction 409 is equivalent to the processing from 402 to 407 in FIG. 2 when the level set algorithm described in the second embodiment is used. In FIG. 6, the function of resetting the parameter of the parameter selector is processed by the parameter reset 410, and the automatic reference data 412 is processed by the automatic reference data updater.

  In FIG. 6, after the lesion part extraction start 400, in the pre-processing 401, as described above, arterial feature amount extraction and parameter selection are processed, and then image extraction 409 is executed. Next, whether or not the result of the image extraction 409 is appropriate is determined in the extraction result approval 411. The determination is performed by the diagnostician. When the determination result is “No”, the diagnostician presents the parameter set by the parameter selection to the diagnostician in the parameter setting 410, and the diagnostician performs appropriate extraction. You can change the parameters as follows. Next, image extraction 409 and extraction result approval 411 are executed again based on the changed parameters. If the determination result is “No”, the changed parameter is presented in the parameter setting 410, and the parameter can be reset as described above. Then, 409 and 411 are repeated.

  On the other hand, when the determination result is “Yes”, the automatic reference data update 412 is processed. In the automatic reference data update 412, the reference data is updated from the feature amount of the artery and the changed parameter that can be appropriately extracted. Thereafter, the extraction ends. Thereby, while operating the said image processing apparatus, reference data can be learned and updated according to environments, such as a diagnostician, and it is not necessary to set reference data beforehand for every operation.

  FIG. 7 shows an example of a screen presented to the diagnostician. FIG. 8 shows an explanatory diagram in which the screen and buttons in FIG. 7 are extracted. 7 and 8 are denoted by the same reference numerals in the equivalent parts. In FIG. 7, an automatic extraction screen 500 is presented to the diagnostician, and a contrast image to be diagnosed is presented in the image display area 501. A control screen is displayed on the right side of the screen 500, a screen 502 for controlling automatic extraction of parameters, and a screen 503 for setting parameters. By selecting the automatic extraction execution button 503, the diagnostician processes the feature extraction and parameter selection of the artery, and presents the parameters in the area of the extraction parameter setting screen 504. The diagnostician can reset the parameters on the screen 504, and can execute the image extraction 409 by selecting the extraction execution button 505. The extraction result is presented on the screen 501 and the diagnostician can execute the determination 411 of the extraction result. When the determination result is “Yes”, the extraction end button 506 is selected.

100: Image processing apparatus,
101: Database server
102_0, 102_1: terminal PC (computer),
103_0, 103_1: display means,
110 — 0,110 — 1: Contrast CT image,
111,111_1: reference data,
120: Arterial feature quantity extractor,
121: Parameter selector,
122: Image filter,
123: Image extractor,
124: post-processor,
125: Reference data updater,
126: Automatic reference data updater,
200: network,
201: Data transfer path of arterial feature amount,
202: Parameter data transfer path for automatic extraction,
203, 204: a network connecting a terminal and a server,
205: Data transfer path for resetting parameters,
300: Reference data projected from the feature amount (horizontal axis) of the artery to the parameter (vertical axis) for extraction,
301: Arterial feature value (horizontal axis) experimentally determined for appropriate extraction and parameters for extraction (vertical axis),
302: a gradation correlation line indicating the correlation between luminance values of an artery and a lesion,
400: Start extraction
401: Pre-processing,
402: Image filter processing,
403: auxiliary function initialization processing,
404: Speed function calculation processing,
405: Auxiliary function update processing,
406: Update amount and threshold comparison processing,
407: post-processing,
408: Extraction finished,
409: image extraction processing,
410: parameter resetting process,
411: Approval processing of extraction results,
412: Automatic reference data update processing,
450: Level set algorithm,
500: Automatic extraction screen,
501: image display area,
502: Control screen for automatic extraction parameter setting;
503: Extraction parameter automatic setting execution button,
504: Extraction parameter setting screen,
505: Automatic extraction execution button,
506: Automatic extraction complete button.

Claims (5)

Input means for capturing medical images;
Image processing means for extracting a lesion from the medical image;
Display means for displaying an image of a lesion extracted by the image processing means,
The medical image is a contrast image obtained by administering a contrast agent to a subject and photographing a measurement object with a tomography apparatus,
Based on the brightness of another tissue different from the lesioned part in the contrast image, the brightness of which changes depending on the contrast agent, set an automatic extraction parameter as a reference when extracting the lesioned part of the subject,
Based on the set the automatic extraction parameters were, we have a means for automatically extracting a lesion from the contrast image,
The other tissue is an artery;
An image processing apparatus , wherein the gradation of the artery is an average value of a histogram of gradation of the artery . The image processing apparatus according to claim 1 .
The processing means includes
An arterial feature extractor that inputs the contrast image and extracts a feature quantity from the gradation of the artery in the image;
A parameter selector for selecting a parameter suitable for automatic extraction of the lesion using the feature amount and reference data which is a reference for setting an automatic extraction parameter;
And an image extractor that extracts a lesion based on the parameter. The image processing apparatus according to claim 2 ,
Automatic extraction of lesions in the image extractor is performed by a level set algorithm,
An image processing apparatus, wherein the automatic extraction parameter is a gradation of a lesioned part. The image processing apparatus according to claim 2 ,
Means for presenting the reference data to a diagnostician;
An image processing apparatus comprising input means for changing the reference data by the diagnostician. The image processing apparatus according to claim 2 ,
An image processing apparatus comprising: means for updating reference data for setting an automatic extraction parameter based on a feature amount output from the arterial feature extractor and a parameter output from the parameter selector.

Images