WO2008038811A1

WO2008038811A1 - Medical image registration system

Info

Publication number: WO2008038811A1
Application number: PCT/JP2007/069130
Authority: WO
Inventors: Tomoaki Gotou; Tetsutarou Ono; Seiji Matsuba
Original assignee: Dai Nippon Printing Co., Ltd.
Priority date: 2006-09-26
Filing date: 2007-09-21
Publication date: 2008-04-03
Also published as: JPWO2008038811A1

Abstract

Provided is a medical image registration system capable of managing an image appropriate for an image processing history. When a medical image is registered into a database, an image processing history of the inputted medical image is compared to an image processing history set as an image registration condition in the database. When they coincide, the medial image is registered unconditionally and when they do not coincide, a message indicating that is displayed.

Description

Description Medical Image Registration System Technical Field

The present invention relates to a database of medical images such as brain images, and more particularly to a technique for registering medical images in a medical image database. Background art

Conventionally, a brain image of a patient is taken with an MR I (Magnetic Resonance Imaging) device and the patient's condition is diagnosed by comparing with a standard image. For example, in the method disclosed in Japanese Patent Application Laid-Open No. 2 025-2 3 7 4 4 1, image processing is performed on brain images of a plurality of healthy persons registered in a database, thereby creating Statistical images such as average images and standard deviation images are used as standard images for comparison. Disclosure of the invention

The brain images of healthy persons registered in the database are processed so that they can be easily compared with other brain images. There are various types of image processing, and different image processing may be performed depending on the equipment or facility that performed the image processing.

When comparing the brain images, if the image processing history of the images to be compared is not known, the images subjected to different image processing may be compared with each other. From such a comparison, there is a problem that a correct diagnosis result cannot be obtained.

Accordingly, an object of the present invention is to provide a medical image registration system capable of managing images suitable for brain image comparison determination.

In order to solve the above problems, a medical image registration system of the present invention is a system for registering a medical image in a database, which is a registration condition of an image registered in the database, and includes an image processing history of the image. Registration condition setting for setting image registration conditions Determining means, image input means for inputting a medical image having image processing history information, image processing history in the image registration condition, and processing history comparison means for comparing the input image processing history information A message output means for outputting a message if the image processing history compared by the processing history comparison means does not match; and an image registration means for registering the medical image input by the image input means in the database. It is characterized by having a configuration having:

In the medical image registration system of the present invention, the images registered in the database are classified into a plurality of categories, and the registration condition setting unit is an image registration condition for each category, An image registration condition including a processing history is set. The image input means has a specifying means for specifying a category of an input image, and the processing history comparison means is specified by the specifying means. The category includes means for comparing the image processing history set by the registration condition setting means with the image processing history inputted by the image input means, and the message output means comprises the processing It has a means for outputting a message when the image processing history in the category compared by the history comparison means does not match. Further, the medical image registration system of the present invention performs statistical processing on all images in the database or the specified category including the new medical image when a new medical image is registered by the image registration unit. It further includes statistical processing image creation means for executing and creating a new statistical processing image.

According to the present invention, when image registration conditions including image processing history are set for a database and a medical image having image processing history information is input, the image processing history set as the image registration conditions in the database and Since the input image processing history is compared and a message is output when the compared image processing histories do not match, medical images that do not match the image processing history may be registered unconditionally. This eliminates inconsistencies in the processing conditions of the images in the database, and has the effect of increasing the reliability of the images to be compared during diagnosis.

Further, according to the present invention, when a new medical image is registered, statistical processing is performed on all the images in the database or the specified category including the new medical image, So that statistically processed images are created. There is an effect that a statistically processed image in a new state can be obtained. Brief Description of Drawings

FIG. 1 is a configuration diagram showing an embodiment of a medical image registration system according to the present invention.

Figure 2 is a diagram showing the relationship between databases and categories and the image registration conditions set for them.

FIG. 3 is a diagram showing an example of feature data used in the present invention.

FIG. 4 is a diagram showing a message display screen displayed and output on the message output means 50. .

FIG. 5 is a diagram showing an example of the image management table in the database 10. FIG. 6 is a diagram showing the relationship between a database and categories and image registration conditions set for them when a detailed image processing history is set.

FIG. 7 is a diagram showing different examples of feature data.

FIG. 8 shows a message display screen when the category image registration conditions do not match. BEST MODE FOR CARRYING OUT THE INVENTION

DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. (1. System configuration)

FIG. 1 is a configuration diagram showing an embodiment of a medical image registration system according to the present invention. In FIG. 1, 10 is a database, 20 is registration condition setting means, 30 is image input means, 40 is processing history comparison means, 50 is message output means, 60 is image registration means, and 70 is It is a statistical processing image creation means.

The database 10 is a database that stores a large number of medical images. The database 10 stores a large number of registered images, and manages these registered images in units of categories. Each category has its image registration conditions set. Further, the database 10 stores statistically processed images obtained by statistically processing registered images accumulated in units of categories. Data such as registered images in database 10 The storage area required for storage is realized by a mass storage device such as a hard disk.

The registration condition setting means 20 sets the image registration conditions for the categories created in order to manage medical images in the entire database 10 and the data base 10. In other words, a common processing history is defined for one database and category, and all medical images belonging to one database and category have undergone the same processing.

The image input means 30 designates a category in the database 10 and inputs a medical image to be registered in the database 10. A medical image is composed of image data itself and feature data including processing history data indicating the processing history. When the medical image to be input is recorded on an external recording medium, the image input means 30 performs a process of reading the medical image from the external recording medium (such as a CD) and uses it as a storage area of the database 10. When a medical image to be input is recorded on the hard disk being used, the medical image is read from the hard disk. The processing history comparison means 40 includes an image processing history included in the feature list of the medical images input by the image input means 30 and an image processing history that is an image registration condition set in the database. Compare the image registration conditions included in the feature data with the image registration conditions set in the category.

The message output means 50 outputs a message when the image processing histories compared by the processing history comparison means 40 do not match. The message output means 50 is realized by a display device such as a liquid crystal display, for example. The image registration means 60 is used to store image data of medical images input by the image input means 30 when the image processing histories compared by the processing history comparison means 40 match. Register in the specified category within 0. The statistical processing image creation means 70, when the new medical image is registered in the specified category in the database 10 by the image registration means 60, includes the newly registered medical image in the category. Statistical processing is performed on all images, and statistical processing images are created. The medical image registration system shown in Fig. 1 is actually realized by executing a dedicated program on a general-purpose computer. (2. System processing operations)

Next, the processing operation of the system shown in FIG. 1 will be described. First, the registration condition setting means 20 sets image registration conditions for the entire database 10. Image processing history is set as the image registration condition for the entire database. As described above, since the image processing history is set as the image registration condition of the database 10, medical images having different image processing histories cannot be registered.

Subsequently, the registration condition setting means 20 creates a category in the database 10 and sets the image registration condition. In the present embodiment, image processing history, age, and gender are set as image registration conditions. For example, when the category “Young Men and Women” is created, the age “40 to 50 years” and the gender “Men / Women” are set in addition to the image processing history as the image registration conditions.

In addition, when the category “elderly man” is created, as the image registration condition, the age “70 to 80 years” and the sex “male” are set in addition to the image processing history. In addition to the image processing history, the age “70 to 80 years” and the gender “female” are set as the image registration conditions.

Figure 2 shows the relationship between the categories created in this way and the set image registration conditions. In this embodiment, the entire database 10 corresponds to the hospital A database, which is one medical facility. In the example of FIG. 2, the image processing history set for the category is omitted.

After the image registration conditions are set, the database 10 is ready for registration of medical images. In this state, the image input means 30 inputs a medical image. Here, the configuration of the medical image will be described. A medical image consists of the image data itself and the feature data.

The image data itself uses the ANAL Y Z E format. This ANAL Y Z E 7 format consists of the entire three-dimensional brain for one person, consisting of a header file and an image file.

Feature data is data that describes the features of the image in text format, and consists of data that describes image processing history, processing parameters, specifications, contents features, and attributes. In this embodiment, the feature data is described in XML format. Figure 3 shows the XM An example of feature data described in L format is shown. The example in Figure 3 shows an example of part of the image processing history and the corresponding processing parameters. In FIG. 3, “resolution conversion” and “tissue separation processing” are described as the image processing history. The resolution conversion processing parameters indicate that the poxel size is 1.2 mm X 1.2 mm X 1.2 mm. The tissue separation processing parameters indicate that the affine conversion is performed. Is shown.

The image processing history indicates the processing performed in the past for the input medical image, and includes resolution conversion, tissue separation processing, smoothing, and anatomical standardization. These specific processing contents are known as disclosed in Japanese Patent Laid-Open No. 2 005-2 3 7 4 4 1, and detailed description thereof is omitted.

The specification in the feature data means, for example, that the data type is 8 bits. In the feature data, the feature of contents means, for example, that the maximum luminance value is within a certain range or that the luminance value force distribution is followed. The attributes in the feature data refer to the age and gender of the person who took the image.

When registering a medical image, a category is specified and the medical image is input from the image input means 30. After inputting the medical image, the processing history comparison means 4 0 first extracts the feature data in the medical image, and then extracts the image processing history in the feature data. The image processing history and the image in the database 10 Compare with registration conditions. If they match, the processing history comparison means 40 further extracts the detailed image processing history, age, and gender from the feature data, and these are specified in the data base 10. Compared with the category image registration conditions of Tegori. If the image registration conditions of the categories match, the input medical image is registered as it is in the specified power category in the data base 10 as it is.

On the other hand, if the information in the feature data of the input medical image does not match either the image registration condition in the database 10 or the category image registration condition, the message output means 50 outputs a message. To do. For example, if the image processing history of the inputted medical image during the feature selection is resolution conversion, tissue separation processing, and smoothing, as shown in Fig. 2, the image registration condition of the A hospital database is resolution conversion , Tissue separation processing, smoothing, anatomical standardization, smoothing, do not match, display message It is powered.

FIG. 4 shows a message display screen displayed on the message output means 50. As shown in Figure 4, the message “The file name rDatall.img of the input medical image does not match the registration conditions of the database” is displayed at the top of the screen. The CT displays the image processing history set in the (image) registration condition of the database and the image processing history described in the feature data of the input medical image. In the example of Fig. 4, five processes are set in the image registration condition of the database, but only three processes are performed on the input medical image. Thus, when the image processing history of the input image does not match the image processing history of the image registration condition in the database, a message is displayed and output.

There are four choices at the bottom of the message display screen, and the operator selects one option. When the operator selects “Register this image”, the medical image is registered even if the image processing history of the medical image does not match the image registration conditions in the database. The lower two of the options are selected by the operator when multiple medical images are input at once. When the operator selects one of the options and clicks the OK button at the bottom, the content selected in the option is executed. As described above, if the image registration conditions to be compared match or the image registration conditions to be compared do not match and the option to register is selected, the input medical image is specified. Registered in the selected category. Here, FIG. 5 shows an example of an image management table in the database 10. As shown in FIG. 5, in the database 10, medical images are managed in units of categories. Each category is associated with a registered medical image. Also, in each category, statistically processed images obtained by statistically processing all medical images in the category are also registered in association with each other. In this embodiment, an average image and a standard deviation image are created as statistically processed images as shown in FIG.

The average image and standard deviation image are images obtained by calculating the average and standard deviation of the pixels at the same position in all images registered in one category and assigning the values as pixel values. In this system, this average image and standard deviation image are recreated each time a new image is registered. Specifically, the image registration means 60 creates a new one. When the medical image is registered in the database 10, the statistical processing image creation means 70 detects the category in which the new medical image is registered, extracts all the images registered in the category, and calculates the average Create images and standard deviation images. Then, the original average image and standard deviation image are deleted, and the newly created average image and standard deviation image are registered in the corresponding category in the database 10.

(3. Processing history by category)

Next, a case where image processing histories are compared for each category will be described. For example, in brain tissue separation processing, separation processing in which the brain is separated into three tissues of gray matter, white matter, and cerebrospinal fluid is often performed. In order to realize such tissue separation processing, information on the distribution probability of the three tissues is necessary, and a standard image of the three tissues is used as a template image. Since these tissues are known to have age-related changes and volume differences depending on gender, the template image is processed more accurately by using an image that matches the characteristics of the subject. Therefore, only the history of the image processing name is specified as the image processing history in the image registration condition of the database, and the parameters of the image processing are specified as the image processing history in the category image registration condition.

In this case, when the registration condition setting means 20 creates the category “Young Men and Women”, as the image registration conditions, the image processing history parameters that are detailed image processing history “Organization Separation Processing, Template Young Men and Women” Set “for”. In addition, when the registration condition setting means 20 creates the category “elderly man”, the image registration condition is changed to “image separation history, template elderly man”, which is a detailed image processing history. When the registration condition setting means 2 0 creates the category “elderly woman”, the image registration parameter “image separation history, template elderly” is used as the image registration condition. "For women" power is set. Figure 6 shows the relationship between the categories created in this way and the image registration conditions that have been set. In this example, the template images used for the tissue separation process, which is the second process, are different for “young boy”, “old man”, and “old woman”.

FIG. 7 shows feature data corresponding to the image registration conditions of FIG. In Fig. 7, the processing parameters for the tissue separation process in the image processing history are “template” and “for young men and women”. This is because tissue separation processing of this image data It is shown that this was done using a template bowl for use.

When a detailed image processing history such as the image processing history parameter is set and a force category is specified and a medical image is input from the image input means 30, the processing history comparison means 40 Compare the image processing history in the data with the image registration conditions in the database 10. If the two match, the processing history comparison means 40 further extracts a detailed image processing history from the feature data, and the category of the category specified in the database 10 and this image processing history. Compare the image registration conditions. If the image registration conditions for the power category also match, the input medical image is registered in the specified category in the database 10 as it is.

On the other hand, if the information in the feature data does not match either the image registration condition in the database or the image registration condition in the category, the message output means 50 displays and outputs a message. For example, if the category “Young Men and Women” is specified and a medical image whose feature data is the image processing history / parameter “Tissue Separation Processing, Template for Elderly Men” is input as the detailed image processing history, as shown in Figure 6. As described above, the image registration condition of the category “Young Men and Women” in the Hospital A database is “Organization Separation Process, Template for Young Men and Women”, which matches the information in the feature data of the input medical image. Message is output.

FIG. 8 shows the state of the message display screen displayed and output by the message output means 50. As shown in Fig. 8, the message “Datal2.img does not match the image registration conditions” is displayed at the top of the screen. Immediately below that, the image processing history set in the image registration conditions and the image processing history described in the feature data of the input medical image are displayed. The image processing history set in the image registration condition is a combination of both the process name of the image registration condition of the database and the processing parameters of the image registration condition of the category. In the example of Fig. 8, the process name set in the image registration condition of the database “1: Resolution conversion” “2: Tissue separation process” “3: Smoothing” “4: Anatomical standardization” “5: “Smoothing” matches the input image processing names “1: Resolution conversion”, “2: Tissue separation processing”, “3: Smoothing”, “4: Anatomical standardization”, and “5: Smoothing”. On the other hand, the tissue separation that is the processing parameter set in the category image registration condition The template for the process is “for young men and women”, and the template for the tissue separation process, which is the processing parameter for the input medical image, is “for elderly men”, and they do not match. As described above, if the image processing history of the input image satisfies the image processing history of the image registration condition of the database, but does not satisfy the image processing history of the category image registration condition, the message is displayed. Is displayed and output.

The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above-described embodiments, and various modifications can be made. For example, in the above embodiment, images of one medical institution are managed on a single database, but by associating one medical institution with each category, one database can be used in a plurality of medical institutions. You may make it manage the image to be managed. In this case, the image processing history is set as an image registration condition of a category corresponding to each medical institution. Industrial applicability

Since the medical image registration system according to the present invention registers a medical image in a database of medical images such as brain images, it can be used in the medical industry and the computer industry.

Claims

The scope of the claims

1. A system for registering medical images in a database,

Registration condition setting means for setting an image registration condition including an image processing history of the image registered in the database,

Image input means for inputting a medical image having image processing history information;

Processing history comparison means for comparing the image processing history in the registration condition with the input image processing history information;

Message output means for outputting a message when the image processing histories compared by the processing history comparison means do not match; and

Image registration means for registering the medical image input by the image input means in the database;

A medical image registration system comprising:

2. The images registered in the database are classified into a plurality of categories, and the registration condition setting means sets image registration conditions including the image processing history, which are image registration conditions for each category. Yes,

The image input means has a designation means for designating a category of an input image, and the processing history comparison means is an image set by the registration condition setting means in the category designated by the designation means. Means for comparing the processing history with the image processing history input by the image input means;

The medical image registration system according to claim 1, wherein the message output means includes means for outputting a message when the image processing histories in the categories compared by the processing history comparison means do not match.

3. When a new medical image is registered by the image registration means, statistical processing is executed for all the images in the database or the specified category including the new medical image, and new statistical processing is performed. 3. The medical image registration system according to claim 1, further comprising statistical processing image creation means for creating an image.

4. A computer is connected to the medical image registration system according to any one of claims 1 to 3. Program to function as a stem.