KR102263708B1 - De-identification device for healthcare image - Google Patents

Abstract

The present invention relates to a de-identification device for a healthcare video using de-identification standard DB. In more detail, the present invention relates to a de-identification device for a healthcare video using de-identification standard DB, which can prevent personal information leakage by using the de-identification standard DB that stores the de-identification standard template information for each specific body part object, and preforming de-identification on personal identification information existing on the healthcare video image or clinical video image, and can provide the original image of a specific patient and the de-identified image at the same time in case of a person who has the right to access to the original video image. The de-identification device includes a video image acquisition part, a personal identification information detection part, a disease area detection part, a de-identification processing part, a de-identification information storage processing part, medical DB, and de-identification standard DB.

Description

De-identification device for healthcare image using de-identification standard DB

The present invention relates to a health care image de-identification device using a de-identification standard DB, and more particularly, to a health care using a de-identification standard DB storing de-identification standard template information for each body specific part object. It prevents personal information leakage by de-identifying the personally identifiable information present on the video image or clinical image image, and provides the original image of a specific patient and the de-identified image at the same time if the person who has access to the original image image is de-identified. It relates to a health care image de-identification device using a de-identification standard DB to make it possible.

Large hospitals with many patients have many medical records and medical images.

And as the 4th industrial environment, the enthusiasm for artificial intelligence using deep learning is increasing, and at the same time, medical imaging research based on it is being actively conducted.

Such medical imaging research is developing and developing with various systems such as deep learning.

It is essential for medical imaging research to secure a certain number of medical images.

In conducting research with various systems such as deep learning, the more samples there are, the more accurate results can be obtained.

Therefore, when researching medical images, medical images are requested as research data from medical institutions that have many medical images, such as large hospitals. will apply

Therefore, it is necessary to manually delete the personal information of the necessary medical images and provide them through the verification process.

This causes a lot of costs, and these costs are emerging as a significant obstacle to research activities due to the increase in research costs.

Therefore, in the present invention, personal identification information existing on a healthcare image image or clinical image image is de-identified by using a de-identification standard DB that stores de-identification standard template information for each object of a specific body part. We propose a device for de-identification of healthcare images using de-identification standard DB to prevent information leakage and to provide the original image and de-identified image of a specific patient at the same time if the person who has access to the original image image it was done

Republic of Korea Patent Publication No. 10-2019-0110498

Accordingly, the present invention has been proposed in view of the problems of the prior art as described above, and a first object of the present invention is to use a de-identification standard DB that stores de-identification standard template information for each body specific part object. It is intended to prevent personal information leakage by de-identifying personal identification information existing on care image or clinical image image.

A second object of the present invention is to provide specific body part objects such as the entire face, eyes, nose, mouth, ears, eyebrows, cheeks, and chin stored in the de-identification standard DB, and specific objects such as tattoos, watches, and name tags that are not stored therein. The de-identification process is different according to the individual recognition object, so that various personal identification information can be de-identified.

A third object of the present invention is to provide an original image and a de-identified image of a specific patient in the case of a person authorized to access the original image.

In order to achieve the problem to be solved by the present invention, the health care image de-identification device using the de-identification standard DB of the present invention,

a video image acquisition unit 100 for obtaining a video image from the outside;

a personal identification information detection unit 200 for detecting a type and area of personal identification information from the obtained video image;

a disease region detection unit 300 for detecting a disease region in the acquired image image;

Converting the image image acquired by the image image acquisition unit 100 according to the type of personal identification information detected by the personal identification information detection unit into a de-identification-processed image image using the de-identification standard DB 700 for de-identification processing unit 400;

The original image image acquired by the image image acquisition unit 100, the type and region of the personal identification information detected by the personal identification information detection unit 200, the disease region detected by the disease region detection unit 300, the ratio a de-identification information storage processing unit 500 for storing and processing the de-identified image converted by the identification processing unit 400 in the medical DB 600;

a medical DB 600 that matches the externally acquired image image original image, the de-identified image, the type and region of personal identification information, and the disease region and stores it for each patient;

Non-identifying standard template information for standard images for each body part (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) by country, age, gender, race, and shooting angle Identification standard DB (700); includes.

Healthcare image de-identification device using de-identification standard DB according to the present invention,

By using the de-identification standard DB that stores de-identification standard template information for each specific body part object, de-identification processing of personal identification information existing on healthcare image images or clinical image images prevents personal information leakage. , it has the effect of eliminating time and economic loss for de-identification processing.

In addition, according to the object of specific body parts such as the entire face, eyes, nose, mouth, ears, eyebrows, cheeks, and chin stored in the de-identification standard DB, and specific personal recognition objects such as tattoos, watches, and name tags that are not stored therein The de-identification processing process is performed differently, so that various personal identification information can be de-identified, thereby providing the effect of processing personal identification types such as various forms and objects.

In addition, by providing the original image and the de-identified image of a specific patient in the case of the person having access to the original image image, the effect of allowing the patient to view the original image of the patient's healthcare and make a more accurate judgment is provided.

1 is a networking conceptual diagram of a healthcare image de-identification device using a de-identification standard DB according to an embodiment of the present invention.
2 is a block diagram of a healthcare image de-identification device using a de-identification standard DB according to an embodiment of the present invention.
3 is a block diagram of a personal identification information detecting unit 200 of a healthcare image de-identification device using a de-identification standard DB according to an embodiment of the present invention.
4 is a block diagram of a de-identification processing unit 400 of a healthcare image de-identification device using a de-identification standard DB according to an embodiment of the present invention.
5 is an exemplary diagram illustrating an eye region and a disease region when the type of personal identification information of the healthcare image de-identification device using the de-identification standard DB according to an embodiment of the present invention is an eye, and FIG. 6 is an image It is an exemplary diagram of detecting a disease region within, and FIG. 7 is an exemplary diagram in which a standard image is taken from the non-identification standard DB 700 and modified to fit the size of the detected region, and FIG. 8 is a disease region added. It is an exemplary view, and FIG. 9 is an exemplary view showing a tattoo area and a diseased area when the detected personal identification information is a tattoo that is an atypical image, FIG. 10 is an exemplary view of detecting a diseased area, and FIG. 11 is a diseased area It is an exemplary view in which the color is changed to the surrounding color, Fig. 12 is an exemplary view in which the detected tattoo area is mosaic-processed, and Fig. 13 is an exemplary view in which a diseased area is added.
14 is a block diagram of an original image provision determining unit for each access identifier of a healthcare image de-identification device using a de-identification standard DB according to an embodiment of the present invention.

The following is merely illustrative of the principles of the invention. Therefore, those skilled in the art will be able to devise various devices which, although not explicitly described or shown herein, embody the principles of the present invention and are included within the spirit and scope of the present invention.

In addition, it should be understood that all conditional terms and examples listed herein are, in principle, expressly intended only for the purpose of understanding the inventive concept and are not limited to the specifically enumerated embodiments and states as such. do.

Healthcare image de-identification device using de-identification standard DB according to an embodiment of the present invention,

a video image acquisition unit 100 for obtaining a video image from the outside;

a personal identification information detection unit 200 for detecting a type and area of personal identification information from the obtained video image;

a disease region detection unit 300 for detecting a disease region in the acquired image image;

Converting the image image acquired by the image image acquisition unit 100 according to the type of personal identification information detected by the personal identification information detection unit into a de-identification-processed image image using the de-identification standard DB 700 for de-identification processing unit 400;

The original image image acquired by the image image acquisition unit 100, the type and region of the personal identification information detected by the personal identification information detection unit 200, the disease region detected by the disease region detection unit 300, the ratio a de-identification information storage processing unit 500 for storing and processing the de-identified image converted by the identification processing unit 400 in the medical DB 600;

a medical DB 600 that matches the externally acquired image image original image, the de-identified image, the type and region of personal identification information, and the disease region and stores it for each patient;

Non-identifying standard template information for standard images for each body part (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) by country, age, gender, race, and shooting angle Identification standard DB (700);

The area of personal identification information detected by the personal identification information detection unit 200 includes a disease area detected by the disease area detection unit 300 .

In addition, according to an additional aspect, the access identifier information, the access identifier for each access identifier authority DB (800) for storing the information on whether or not to provide the original medical DB image for each access identifier;

When the access terminal 3000 is connected, in the case of a visitor who is permitted to provide the original image, the de-identified image of the patient stored in the medical DB 600 and the original image are simultaneously provided to the access terminal 3000 by the access terminal 3000 . It is characterized in that it further comprises; the original image provision decision unit 900 for each access identifier.

At this time, the original image provision determination unit 900 for each connection identifier,

Providing the authorization setting page for receiving the original image to the administrator terminal, and storing and processing information on whether to provide the original image by access identifier set on the authorization setting page provided by the administrator terminal in the authorization DB 800 for each access identifier an original image provision permission setting module 910 for;

When the access terminal 3000 is connected, the access identifier extraction module 920 for extracting the access identifier information from the access information;

Based on the extracted access identifier information, a visitor authority determination module for determining whether the corresponding visitor is a visitor permitted to provide the original image by comparing it with the information on whether or not to provide the original image for each access identifier stored in the permission DB 800 for each access identifier ( 930);

As a result of determination by the accessor's authority determination module 930, if the accessor is permitted to provide the original image, the de-identified image and the original image of the patient stored in the medical DB 600 with the corresponding access terminal 3000 are An original image providing module 940 for providing; characterized in that it is configured to include.

In addition, the image image acquired by the image image acquisition unit 100 is a healthcare image image or a clinical image image.

In this case, the personal identification information is characterized in that it includes object information on specific parts of the body such as the entire face, eyes, nose, mouth, ears, eyebrows, cheeks, and chin, and information on specific personal recognition objects such as tattoos, watches, and name tags.

In addition, the personal identification information detection unit 200,

a first personal identification type detection module 210 for detecting the type and area of personal identification information by comparing it with the artificial intelligence learning result based on the shape information for each part analyzed with the characteristic points of the human body;

In the case of a type of personal identification information not detected by the first personal identification type detection module, it is defined as a non-standard type, and a second personal identification type detection module 220 for detecting the corresponding area; characterized in that it includes; .

In addition, the disease region detection unit 300 is characterized in that it detects the type and region of the disease by comparing it with the artificial intelligence learning result based on the disease shape information.

In addition, the de-identification processing unit 400,

a standard image determination module 410 for determining whether the type of personal identification information detected by the personal identification information detection unit 200 is information stored in the de-identification standard DB 700;

As a result of determination by the standard image determination module 410, if the information is stored in the non-identification standard DB 700, the image of the disease region detected by the disease region detection unit 300 is extracted from the acquired image image. a first disease region acquisition module 420;

Remove the area of personal identification information detected by the personal identification information detection unit 200 from the acquired image image, extract a standard image corresponding to the type of personal identification information from the de-identification standard DB 700, and extract the standard image After adjusting the size of the to fit the size of the area of the removed personal identification information, the standard image whose size has been adjusted is inserted into the area removed on the acquired image image from which the area of personal identification information has been removed to generate a first replacement image. standard image replacement module 430;

The image of the disease region extracted and acquired by the first disease region acquisition module 420 is inserted into the disease region in the first replacement image generated by the standard image replacement module 430 to generate a de-identified image. and a first de-identification processing module (440).

In addition, according to an additional aspect, the de-identification processing unit 400,

As a result of determination by the standard image determination module 410, if it is not the information stored in the non-identification standard DB 700, the image of the disease region detected by the disease region detection unit 300 is extracted and obtained from the acquired image image a second disease region acquisition module 450 for;

a disease region color conversion module 460 for removing the disease region detected by the disease region detection unit 300 from the acquired image image and converting the color of the removed disease region into the same color as the surrounding color;

Mosaic processing module 470 for generating a second replacement image by mosaic processing the area of the personal identification information detected by the personal identification information detection unit 200 on the acquired image image in which the color of the disease region is changed to the surrounding color;

The image of the disease region extracted and acquired by the second disease region acquisition module 450 is inserted into the disease region in the second replacement image generated by the mosaic processing module 470 to generate a de-identified image image. The second de-identification processing module 480; characterized in that it further comprises.

In addition, the standard image replacement module 430 applies a face recognition algorithm when extracting a standard image from the de-identification standard DB 700, and a standard image showing a matching rate of more than a certain probability with an image corresponding to the type of personal identification information. characterized in that it is extracted.

In addition, the medical DB 600 is at least one of PACS (Picture Archiving and Communication System), EMR (Electronic Medical Record), PHR (Personal Health Record), RIS (Radiology Information System), or a cloud server that can be connected thereto. characterized.

Hereinafter, it will be described in detail through an embodiment of a healthcare image de-identification device using a de-identification standard DB according to the present invention.

1 is a networking conceptual diagram of a healthcare image de-identification device using a de-identification standard DB according to an embodiment of the present invention.

As shown in FIG. 1, the healthcare image de-identification device using the de-identification standard DB is an individual existing on a healthcare image image or a clinical image image acquired from an external image acquisition device 2000 connected through a network. The identification information is de-identified based on the de-identification standard DB 700 to perform a function of preventing personal information leakage.

In addition, when a specific connection terminal 3000 is connected, the de-identified image and the original image of the patient stored in the medical DB 600 are stored in the medical DB 600 with the connection terminal 3000 if the user is permitted to provide the original image. function will be provided.

The external image acquisition device 2000 may be a health care image image or a photographing apparatus for photographing clinical image images, for example, a clinical image photographing apparatus, various smart phones, a health care photographing apparatus, a dermascope apparatus, and the like. The health care image images are transmitted to the health care image de-identification device 1000 through the network.

2 is a block diagram of a healthcare image de-identification device using a de-identification standard DB according to an embodiment of the present invention.

As shown in FIG. 2, the apparatus for de-identification of healthcare images using the de-identification standard DB of the present invention is,

a video image acquisition unit 100 for obtaining a video image from the outside;

a personal identification information detection unit 200 for detecting a type and area of personal identification information from the obtained video image;

a disease region detection unit 300 for detecting a disease region in the acquired image image;

Converting the image image acquired by the image image acquisition unit 100 according to the type of personal identification information detected by the personal identification information detection unit into a de-identification-processed image image using the de-identification standard DB 700 for de-identification processing unit 400;

The original image image acquired by the image image acquisition unit 100, the type and region of the personal identification information detected by the personal identification information detection unit 200, the disease region detected by the disease region detection unit 300, the ratio a de-identification information storage processing unit 500 for storing and processing the de-identified image converted by the identification processing unit 400 in the medical DB 600;

a medical DB 600 that matches the externally acquired image image original image, the de-identified image, the type and region of personal identification information, and the disease region and stores it for each patient;

Non-identifying standard template information for standard images for each body part (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) by country, age, gender, race, and shooting angle Identification standard DB (700); is configured to include.

Specifically, the video image acquisition unit 100 acquires a video image from the outside. That is, as shown in FIGS. 5 to 13 , any one of a healthcare image image or a clinical image image is acquired from an external image acquisition device 2000 connected through a network.

As described in the background art, the video images acquired from the outside for diagnosis include personal information that can infer a specific individual upon exposure, and may be, for example, a special eye shape, a special nose shape, or a special tattoo.

If, after diagnosis, it is stored as it is, a problem of personal information leakage may occur, and thus the device of the present invention is proposed to prevent this.

The personal identification information detection unit 200 includes the type of personal identification information (eg, the entire face, eyes, nose, mouth, tattoo, watch, name tag, etc.) in the obtained image image (health care image image or clinical image image) and It performs a function to detect an area.

For example, if the acquired video image is the same as in FIG. 5 (when a disease occurs under the eyes), the eye is the type of personal identification information, and the area of the eye (10 in FIG. 5) is detected as the area of personal identification information. will become

The personal identification information described in the present invention is characterized in that it is object information of specific parts of the body such as the entire face, eyes, nose, mouth, ears, eyebrows, cheeks, and chin, and specific personal recognition object information such as tattoos, watches, and name tags. do.

In order to perform the above function, as shown in Figure 3, the personal identification information detection unit 200,

A video image obtained by comparing the learning results obtained by artificial intelligence based on the shape information of each part (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) analyzed with the characteristic points of the human body and the obtained video image a first personal identification type detection module 210 for detecting the type and area of personal identification information;

a second personal identification type detection module 220 that defines a non-standard type when the personal identification information is a type of personal identification information not detected by the first personal identification type detection module 210 and detects a corresponding area; It is characterized in that it includes.

More specifically, the first individual identification type detection module 210 AI-learns shape information for each part (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) analyzed as characteristic points of the human body. It performs a function of detecting the type and area of personal identification information by comparing the learning result with the acquired video image.

For example, feature points of the human body's eyes, nose, mouth, eyebrows, cheeks, entire face, and chin are collected to provide shape information for each part.

In addition, the artificial intelligence algorithm of the present invention is an algorithm that performs shape learning based on the shape information.

Accordingly, the image acquired by the image image acquisition unit 100 and the learning result of artificial intelligence learning shape information for each part (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) analyzed with the characteristic points of the human body By comparing the images, it is possible to determine which body part the video image is, and it is possible to identify the identified body part region.

The artificial intelligence algorithm using shape learning is a generalized technology and a detailed description thereof will be omitted.

The second personal identification type detection module 220 defines a non-standard type when the personal identification information is a type of personal identification information that is not detected by the first personal identification type detection module, and functions to detect the corresponding area. will perform

For example, as shown in FIG. 9 , in the case of a video image of a skin disease occurring near the tattooed skin, the tattoo is of a non-standard type (eyes, nose, mouth, eyebrows, cheeks, entire face, chin of the human body) not a body part such as the back).

Accordingly, in the case of a tattoo, since it corresponds to the type of personal identification information not detected by the first personal identification type detection module, the second personal identification type detection module 220 defines the area as a non-standard type and detects the area. it will be done

The disease region detection unit 300 performs a function for detecting a disease region in the acquired image image.

For example, as shown in FIG. 5 , the disease region 20 is detected in the acquired video image.

The disease region detection unit 300 detects the type and region of a disease existing in the acquired image image by comparing the acquired image image with the learning result of artificial intelligence learning based on the disease shape information.

That is, after learning the types of various disease types through artificial intelligence learning, the type of disease is determined by comparing the learning result with the acquired image image, and the region of the disease is detected.

The artificial intelligence algorithm using shape learning for each type of disease is a generalized technology, and a detailed description thereof will be omitted.

In particular, as shown in FIG. 5 , the area of personal identification information detected by the personal identification information detecting unit 200 includes a disease area detected by the disease area detecting unit 300 .

The de-identification processing unit 400 de-identifies the image image acquired by the image image acquisition unit 100 according to the type of personal identification information detected by the personal identification information detection unit using the de-identification standard DB 700 . It performs the function to convert the video image that has been processed.

That is, depending on the type of personal identification information, the image image acquired by the image image acquisition unit 100 (eg, an image including images of eyes, nose, mouth, etc.) is a predefined standard image image or not defined. The de-identification process is changed according to the non-standard image image, which is then converted into a de-identified image.

To this end, as shown in Figure 4, the de-identification processing unit 400,

a standard image determination module 410 for determining whether the type of personal identification information detected by the personal identification information detection unit 200 is information stored in the de-identification standard DB 700;

As a result of determination by the standard image determination module 410, if it is information stored in the non-identification standard DB 700, the image of the disease region detected by the disease region detection unit 300 is extracted from the acquired image image. a first disease region acquisition module 420;

Remove the area of personal identification information detected by the personal identification information detection unit 200 from the acquired image image, extract a standard image corresponding to the type of personal identification information from the de-identification standard DB 700, and extract the standard image After adjusting the size of the to fit the size of the area of the removed personal identification information, the standard image whose size has been adjusted is inserted into the area removed on the acquired image image from which the area of personal identification information has been removed to generate a first replacement image. standard image replacement module 430;

The image of the disease region extracted and acquired by the first disease region acquisition module 420 is inserted into the disease region in the first replacement image generated by the standard image replacement module 430 to generate a de-identified image. and a first de-identification processing module (440).

Specifically, the standard image determination module 410 performs a function of determining whether the type of personal identification information detected by the personal identification information detection unit 200 is information stored in the de-identification standard DB 700 .

For example, in the case of FIG. 5 , the type of personal identification information corresponds to the eye, and it is determined whether the image information 'eye' is stored in the standard image DB 700 .

As a result of determination by the standard image determination module 410, the first disease region acquisition module 420 determines the disease region detected by the disease region detection unit 300 in the case of information stored in the non-identification standard DB 700. A function of extracting and acquiring an image from the acquired video image is performed.

For example, as shown in FIG. 6 , the image of the disease region 20 detected by the disease region detection unit 300 is extracted and acquired from the acquired image image.

Thereafter, the standard image replacement module 430 removes the area of personal identification information detected by the personal identification information detection unit 200 from the acquired image image, and de-identifies the standard image corresponding to the type of personal identification information. After extracting from the DB 700 and adjusting the size of the extracted standard image to the size of the area of the removed personal identification information, the size-adjusted standard image is removed from the acquired image image from which the area of personal identification information has been removed The function of generating the first replacement image is performed by inserting it into the region.

For example, the eye area (10 shown in FIG. 5) corresponding to the area of personal identification information is removed from the acquired image image, and the eye standard image corresponding to the type of personal identification information is removed from the de-identification standard DB ( 700), the size of the extracted standard image is adjusted to the size of the removed personal identification information area (adjusted to the size of the dotted line area in FIG. 7), and then the size of the standard eye image (standard eye template image) A first replacement image is generated by inserting the .

On the other hand, the standard image replacement module 430 applies a face recognition algorithm when extracting a standard image from the de-identification standard DB 700, and a standard image showing a matching rate of more than a certain probability with an image corresponding to the type of personal identification information. characterized in that it is extracted.

For example, if the predetermined probability is set to 80%, when the first replacement image is generated, the standard image to be inserted into the removal area on the video image from which the area of personal identification information has been removed is the image for the removed personal identification information ( Example: Removed eye image) and standard image with 80% or higher matching rate (removed eye image and standard eye image with 80% or higher matching rate) are extracted from de-identification standard DB 700 using face recognition algorithm will do

The de-identification standard DB 700 is a ratio of standard images for each part of the human body (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) by country, age, gender, race, and shooting angle. It is characterized in that the identification standard template information is stored.

For example, standard image information of each part of the human body by country, age, gender, race, and shooting angle, such as the entire face, eyes, nose, mouth, ears, eyebrows, cheeks, and chin, specific body parts (standard template information) is stored, so that the above-described standard image replacement module 430 can be used as a replacement image.

The reason for storing de-identification standard template information about standard images for each part of the human body (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) for each shooting angle in the de-identification standard DB 700 is This is to create a natural first alternative image by using images for each part of the human body at various angles.

Then, the first de-identification processing module 440 converts the image of the disease region extracted and acquired by the first disease region acquisition module 420 to the corresponding disease in the first replacement image generated by the standard image replacement module 430 . It performs the function of generating the de-identified image image by inserting it into the area.

For example, as shown in FIG. 8 , a disease extracted and acquired by the first disease region acquisition module 420 in the corresponding disease region on the first alternate image ( FIG. 7 ) generated by the standard image replacement module 430 . By inserting an image of a region, a de-identified video image is generated.

On the other hand, according to an additional aspect, the de-identification processing unit 400,

As a result of determination by the standard image determination module 410, if it is not the information stored in the non-identification standard DB 700, the image of the disease region detected by the disease region detection unit 300 is extracted and obtained from the acquired image image a second disease region acquisition module 450 for;

a disease region color conversion module 460 for removing the disease region detected by the disease region detection unit 300 from the acquired image image and converting the color of the removed disease region into the same color as the surrounding color;

Mosaic processing module 470 for generating a second replacement image by mosaic processing the area of the personal identification information detected by the personal identification information detection unit 200 on the acquired image image in which the color of the disease region is changed to the surrounding color;

The image of the disease region extracted and acquired by the second disease region acquisition module 450 is inserted into the disease region in the second replacement image generated by the mosaic processing module 470 to generate a de-identified image image. The second de-identification processing module 480; characterized in that it further comprises.

Specifically, as a result of the determination by the standard image determination module 410, the second disease region acquisition module 450 determines that the type of personal identification information detected by the personal identification information detection unit 200 is a de-identification standard DB ( If it is not the information stored in 700), the function for extracting and acquiring the image of the diseased region detected by the disease region detection unit 300 from the acquired image image is performed.

For example, as shown in FIG. 9 , when the acquired image image is an image in which the skin disease occurs next to the tattoo, the type of personal identification information detected by the personal identification information detection unit 200 is the de-identification standard. If it is not information stored in the DB 700 (a tattoo is not information about an object of a specific body part such as the entire face, eyes, nose, mouth, ears, eyebrows, cheeks, and chin), the disease area detection unit 300 detects An image of the diseased region (30 in FIG. 10) is extracted and acquired from the acquired image image.

Thereafter, the disease region color conversion module 460 removes the disease region detected by the disease region detection unit 300 from the acquired image image, and converts the color of the removed disease region into the same color as the surrounding color. will do For example, the diseased area (30 in FIG. 10) is removed and the color of the removed area (30 in FIG. 11) is converted into a skin color that is the same color as the surrounding color.

Thereafter, the mosaic processing module 470 mosaics the area of the personal identification information detected by the personal identification information detection unit 200 among the areas on the acquired image image in which the color of the diseased area is changed to the surrounding color to generate a second replacement image. function to create.

For example, as shown in FIG. 12 , the personal identification information is de-identified by mosaic processing the area of the personal identification information detected by the personal identification information detection unit 200 ( 40 in FIG. 12 ).

In this case, in the case of a unique tattoo, it is difficult to recognize whose tattoo it is because it is mosaic-treated.

Thereafter, the second de-identification processing module 480 converts the image of the disease region extracted and acquired by the second disease region acquisition module 450 to the corresponding disease region on the second replacement image generated by the mosaic processing module 470 . It performs the function of generating a de-identified video image by inserting it into the .

For example, as shown in FIG. 13 , the disease region image 30 extracted and acquired by the second disease region acquisition module 450 is the disease region on the second replacement image generated by the mosaic processing module 470 . It is to create a de-identified video image by inserting it into

The de-identification information storage processing unit 500 includes the original image image acquired by the image image acquisition unit 100, the type and area of personal identification information detected by the personal identification information detection unit 200, and the disease area detection unit The disease region detected by 300 and the de-identified image converted by the de-identification processing unit 400 are stored and processed in the medical DB 600 .

For example, the de-identified image as shown in FIG. 8 and the type of personal identification information such as eyes, eye regions, and disease regions located under the eyes are stored and processed in the medical DB 600 .

On the other hand, the medical DB 600 described in the present invention is at least one of PACS (Picture Archiving and Communication System), EMR (Electronic Medical Record), PHR (Personal Health Record), RIS (Radiology Information System), or connected thereto. It is characterized as a possible cloud server.

Through this, the specialists access the medical DB 600 and diagnose the patient's skin disease through the unidentified image image of the specific patient, but identify the patient's personal information based on the characteristic body part in the diagnosis process. it will be impossible

In addition, the medical DB 600 matches the source image of the externally acquired image, the de-identified image, the type and region of personal identification information, and the disease region, and stores it for each patient.

That is, the original image image shown in FIG. 5, the de-identified image image shown in FIG. 8, and the types of personal identification information such as eyes, eye regions, and disease regions located under the eyes are matched and stored for each patient. .

On the other hand, according to an additional aspect, the health care image de-identification device using the de-identification standard DB of the present invention,

Access identifier information, access identifier for each access identifier authority DB (800) for storing the information on whether or not to provide the original image of the medical DB for each access identifier;

When the access terminal 3000 is connected, in the case of a visitor who is permitted to provide the original image, the de-identified image of the patient stored in the medical DB 600 and the original image are simultaneously provided to the access terminal 3000 by the access terminal 3000 . It is characterized in that it further comprises; the original image provision decision unit 900 for each access identifier.

More specifically, the access identifier information and information on whether to provide the original medical DB image for each access identifier are stored in the access identifier-specific authority DB 800 .

For example, access identifier information is given to each health care worker named Doctor A: A0101 and Doctor B: B0111, and the right to provide the original image when accessing the medical DB for each access identifier is stored.

When the access terminal 3000 is connected, the original image provision determining unit 900 for each access identifier is the access terminal 3000, when the access terminal is permitted to provide the original image, the ratio of the patient stored in the medical DB 600 The identifiable image and the original image are simultaneously provided. The access terminal 3000 means a terminal of a medical practitioner (specialist).

For example, when a specialist A having an access identifier of A0101 accesses using the access terminal 3000, it can be known that it is a specialist A, and if the authority to provide the original image is granted to the specialist A, it is shown in FIG. A de-identified image and the original image shown in FIG. 5 can be simultaneously provided.

On the other hand, when a specialist B having an access identifier of B0111 accesses using the access terminal 3000, it is known that he is a specialist B, and if the authority to provide the original image is not granted to the specialist B, Only the identified image is provided, and the original image shown in FIG. 5 is not provided.

As described above, if the original image is provided to the authorizer, it is possible to prevent leakage of sensitive personal information, and accordingly, it is possible to inspire the need to protect personal information.

In order to provide the above functions, as shown in FIG. 14, the original image provision determining unit 900 for each connection identifier,

Providing the authorization setting page for receiving the original image to the administrator terminal, and storing and processing information on whether to provide the original image by access identifier set on the authorization setting page provided by the administrator terminal in the authorization DB 800 for each access identifier an original image provision permission setting module 910 for;

a connection identifier extraction module 920 for extracting the connection identifier information of the corresponding contact from the access information when the access terminal 3000 is connected;

Based on the extracted access identifier information, a visitor authority determination module for determining whether the corresponding visitor is a visitor permitted to provide the original image by comparing it with the information on whether or not to provide the original image for each access identifier stored in the permission DB 900 for each access identifier ( 930);

As a result of determination by the accessor's authority determination module 930, if the accessor is permitted to provide the original image, the de-identified image and the original image of the patient stored in the medical DB 600 with the corresponding access terminal 3000 are An original image providing module 940 for providing; characterized in that it is configured to include.

More specifically, the original image provision permission setting module 910 provides an authorization setting page for receiving the original image stored in the medical DB 600 to the administrator terminal.

In this case, the authorization setting is provided to the administrator terminal so that the administrator can set it, such as the final person in charge of the general affairs department or the head of the hospital, and the administrator can grant the authorization.

In addition, information on whether to provide original images for each access identifier set on the authorization setting page provided by the administrator terminal is stored in the authorization DB 800 for each access identifier.

For example, on the authorization setting page, the administrator sets the information on whether or not to provide the original image: A specialist: A0101-authorization approval, B specialist: B0111-non-approval, which is set by the original image provision permission setting module 910 is stored and processed in the permission DB 800 for each access identifier.

When the access terminal 3000 is connected, the access identifier extraction module 920 performs a function for extracting the access identifier information of the corresponding contact from the access information.

For example, when Specialist A accesses using the access terminal 3000, he/she inputs access information such as ID and password, and uses the inputted access information to connect, for example, Doctor A: A0101. Identifier information is extracted.

The accessor authority determining module 930 compares the information on whether or not the original video is provided for each access identifier stored in the authority DB 800 for each access identifier based on the extracted access identifier information to determine whether the corresponding accessor is the accessor who is permitted to provide the original video. will judge

For example, if the access identification letter information is Specialist A: A0101, and the information on whether or not to provide the original image by access identifier stored in the permission DB 800 for each access identifier is Specialist A: A0101-authorization approval, the corresponding visitor provides the original image It will be judged as an allowed accessor.

Thereafter, as a result of determination by the accessor's authority determination module 930, the original image providing module 940 uses the corresponding access terminal 3000 to provide the original image of the corresponding patient stored in the medical DB 600 when the accessor is permitted to provide the original image. The de-identified image and the original image will be provided.

For example, by providing the de-identified image image shown in FIG. 8 and the original image image information shown in FIG. 5 with the connection terminal 3000 accessed by the doctor A, the doctor A can comprehensively diagnose the disease. to be able to judge.

According to the present invention, personal identification information existing on a healthcare image image or a clinical image image is de-identified by using a de-identification standard DB that stores de-identification standard template information for each object of a specific body part. By preventing information leakage, there is an effect of eliminating time and economic loss for de-identification processing.

In addition, by providing the original image and the de-identified image of a specific patient in the case of the person having access to the original image image, the effect of allowing the patient to view the original image of the patient's healthcare and make a more accurate judgment is provided.

In the above, preferred embodiments of the present invention have been illustrated and described, but the present invention is not limited to the specific embodiments described above, and it is common in the technical field to which the present invention pertains without departing from the gist of the present invention as claimed in the claims. Various modifications may be made by those having the knowledge of, of course, and these modifications should not be individually understood from the technical spirit or perspective of the present invention.

100: video image acquisition unit
200: personally identifiable information detection unit
300: disease area detection unit
400: de-identification processing unit
500: de-identified information storage processing unit
600: medical DB
700: De-identified standard DB
800: Authority DB for each access identifier
900: original image provision judging unit for each access identifier

Claims (11)

In the healthcare image de-identification device using the de-identification standard DB,
a video image acquisition unit 100 for obtaining a video image from the outside;
a personal identification information detection unit 200 for detecting a type and area of personal identification information from the obtained video image;
a disease region detection unit 300 for detecting a disease region in the acquired image image;
Converting the image image acquired by the image image acquisition unit 100 according to the type of personal identification information detected by the personal identification information detection unit into a de-identification-processed image image using the de-identification standard DB 700 de-identification processing unit 400 for;
The original image image acquired by the image image acquisition unit 100, the type and region of the personal identification information detected by the personal identification information detection unit 200, the disease region detected by the disease region detection unit 300, the ratio a de-identification information storage processing unit 500 for storing and processing the de-identified image converted by the identification processing unit 400 in the medical DB 600;
a medical DB 600 that matches the externally acquired image image original image, the de-identified image, the type and region of personal identification information, and the disease region and stores it for each patient;
Non-identifying standard template information for standard images for each body part (full face, eyes, nose, mouth, eyebrows, cheeks, chin, etc.) by country, age, gender, race, and shooting angle Identification standard DB (700);
The area of personal identification information detected by the personal identification information detection unit 200 includes a disease area detected by the disease area detection unit 300,

The de-identification processing unit 400,
a standard image determination module 410 for determining whether the type of personal identification information detected by the personal identification information detection unit 200 is information stored in the de-identification standard DB 700;
As a result of determination by the standard image determination module 410, if it is information stored in the non-identification standard DB 700, the image of the disease region detected by the disease region detection unit 300 is extracted from the acquired image image. a first disease region acquisition module 420;
Remove the area of personal identification information detected by the personal identification information detection unit 200 from the acquired image image, extract a standard image corresponding to the type of personal identification information from the de-identification standard DB 700, and extract the standard image After adjusting the size of the to fit the size of the area of the removed personal identification information, the standard image whose size has been adjusted is inserted into the area removed on the acquired image image from which the area of personal identification information has been removed to generate a first replacement image. standard image replacement module 430;
The image of the disease region extracted and acquired by the first disease region acquisition module 420 is inserted into the disease region in the first replacement image generated by the standard image replacement module 430 to generate a de-identified image. A health care image de-identification device using a de-identification standard DB, characterized in that it comprises a first de-identification processing module (440).
The method of claim 1,
Access identifier information, access identifier for each access identifier authority DB (800) for storing the information on whether or not to provide the original image of the medical DB for each access identifier;
When the access terminal 3000 is connected, in the case of a visitor who is permitted to provide the original image, the de-identified image of the patient stored in the medical DB 600 and the original image are simultaneously provided to the access terminal 3000 by the access terminal 3000 . Health care image de-identification device using de-identification standard DB, characterized in that it further comprises;
3. The method of claim 2,
The original image provision determination unit 900 for each access identifier,
Providing the authorization setting page for receiving the original image to the administrator terminal, and storing the information on whether to provide the original image by access identifier set on the authorization setting page provided by the administrator terminal in the authorization DB 800 for each access identifier an original image providing permission setting module 910 for;
When the access terminal 3000 is connected, the access identifier extraction module 920 for extracting the access identifier information from the access information;
Based on the extracted access identifier information, a visitor authority determination module for determining whether the corresponding visitor is a visitor permitted to provide the original image by comparing it with the information on whether or not to provide the original image for each access identifier stored in the permission DB 800 for each access identifier ( 930);
As a result of determination by the accessor's authority determination module 930, if the accessor is permitted to provide the original image, the de-identified image and the original image of the patient stored in the medical DB 600 with the corresponding access terminal 3000 are Health care image de-identification device using a de-identification standard DB, characterized in that it comprises; an original image providing module 940 for providing.
The method of claim 1,
The video image acquired by the video image acquisition unit 100 is,
Healthcare image de-identification device using de-identification standard DB, characterized in that it is a health care image image or a clinical image image.
The method of claim 1,
The personally identifiable information is
Healthcare using de-identification standard DB, characterized in that it includes object information on specific parts of the body such as the entire face, eyes, nose, mouth, ears, eyebrows, cheeks, and chin, and specific personal recognition object information such as tattoos, watches, and name tags Image de-identification device.
The method of claim 1,
The personal identification information detection unit 200,
a first personal identification type detection module 210 for detecting the type and area of personal identification information by comparing it with the artificial intelligence learning result based on the shape information for each part analyzed with the characteristic points of the human body;
and a second personal identification type detection module 220 for defining a non-standard type when the type of personal identification information not detected by the first personal identification type detection module and detecting the corresponding area; Healthcare image de-identification device using de-identification standard DB.
The method of claim 1,
The disease region detection unit 300,
A health care image de-identification device using a de-identification standard DB, characterized in that it detects the type and area of a disease by comparing it with the result of artificial intelligence learning based on disease shape information
delete The method of claim 1,
The de-identification processing unit 400,
As a result of determination by the standard image determination module 410, if it is not the information stored in the non-identification standard DB 700, the image of the disease region detected by the disease region detection unit 300 is extracted and obtained from the acquired image image a second disease region acquisition module 450 for;
a disease region color conversion module 460 for removing the disease region detected by the disease region detection unit 300 from the acquired image image and converting the color of the removed disease region into the same color as the surrounding color;
Mosaic processing module 470 for generating a second replacement image by mosaic processing the area of the personal identification information detected by the personal identification information detection unit 200 on the acquired image image in which the color of the disease region is changed to the surrounding color;
The image of the disease region extracted and acquired by the second disease region acquisition module 450 is inserted into the disease region in the second replacement image generated by the mosaic processing module 470 to generate a de-identified image image. A health care image de-identification device using a de-identification standard DB, characterized in that it further comprises a second de-identification processing module (480).
The method of claim 1,
The standard image replacement module 430,
When the standard image is extracted from the de-identification standard DB (700), the de-identification standard characterized in that the image corresponding to the type of personal identification information and the standard image showing a matching rate of more than a certain probability are extracted by applying a face recognition algorithm Healthcare image de-identification device using DB.
The method of claim 1,
The medical DB 600,
At least one of PACS (Picture Archiving and Communication System), EMR (Electronic Medical Record), PHR (Personal Health Record), RIS (Radiology Information System), or a cloud server that can be connected thereto. Health care image de-identification device used.

Images