KR102707242B1 - System for de-identifying images containing personal information in documents and method thereof - Google Patents

Abstract

The present invention relates to a system and method for de-identifying images containing personal information in a document, which can prevent exposure of personal information contained in images in a PDF file by extracting images in various formats (i.e., TIFF, JPEG, PNG, and BMP) in a PDF file, generating de-identified images according to the corresponding image formats, and updating the de-identified images generated according to each image format in an original PDF file.

Description

{SYSTEM FOR DE-IDENTIFYING IMAGES CONTAINING PERSONAL INFORMATION IN DOCUMENTS AND METHOD THEREOF}

The present invention relates to a system and method for de-identifying images containing personal information contained in a document, and more particularly, to a system and method for de-identifying images containing personal information contained in a document, which extracts images in various formats (i.e., TIFF, JPEG, PNG, and BMP) contained in a PDF file, generates de-identified images according to the corresponding image formats, and updates the de-identified images generated according to each image format to the original PDF file, thereby preventing exposure of personal information contained in images contained in a PDF file.

As the need for the use of personal information increases in cutting-edge industries such as artificial intelligence and self-driving cars, as well as in online shopping malls that provide customized customer services, the demand for establishing a system for anonymizing personal information to protect personal information while enabling its smooth use is rapidly increasing.

In addition, the government allows the collection, use, and provision of personal information as a priority in situations where it is deemed urgently necessary for the protection of the lives, bodies, and property interests of citizens, public health, and other public safety and well-being, but requires compliance with regulations on safety measures, destruction, and the rights of information subjects.

For example, de-identification of personal information according to the Personal Information Protection Act means taking measures to make a specific individual unidentifiable by deleting or modifying part or all of the personal information. As specific de-identification measures, the domestic guideline presents 17 detailed technologies categorized into a total of 5 processing techniques: pseudonymization including heuristic pseudonymization, encryption, and exchange methods; aggregation including total processing, partial aggregation, rounding, and rearrangement; data reduction including identifier deletion, partial identifier deletion, record deletion, and deletion of all identifying elements; data suppression including hiding, random rounding, range method, and controlled rounding; and data masking including adding random noise and blanking and substitution.

The personal information filtering solutions currently in use filter the text contained in the document to provide information that specific personal information or sensitive information is included. In addition, the PDF file format, which is an international standard document, can create a new version by removing or modifying the text if it contains personal information or sensitive information.

However, the masking and de-identification methods used on various image format documents included in PDF files were limited to simply covering some areas of the image with various geometric objects such as circles and squares.

In addition, since the existing PDF black masking or de-identification solution products adopt the same method as the 17 detailed technologies classified into the 5 processing techniques described above, the information hidden by inserting another geometric object over the image object can be easily removed using a commonly used PDF editing tool, which can be said to be at serious risk of personal information exposure.

Therefore, the present invention proposes a method for preventing exposure of personal information contained in images within a PDF file by anonymizing personal information contained in images in various image formats, such as TIFF, JPEG, PNG, and BMP, within a PDF file.

Next, prior inventions existing in the technical field of the present invention will be briefly explained, and then technical matters that the present invention seeks to achieve differently from the prior inventions will be described.

First, Korean Patent No. 10-2319492 (October 29, 2021) is a prior invention regarding a personal information processing system using AI deep learning and a personal information processing method using the same that effectively detects and de-identifies personal information held by individuals or organizations by analyzing text and image information contained in an image file using an artificial intelligence neural network and then detecting and extracting personal information (such as personal identification number information or fingerprint information) contained therein.

In addition, Korean Patent No. 10-2585793 (October 6, 2023) is a prior invention regarding a medical image processing method that generates a safe image in which the risk information is not identifiable in a medical image by changing the color of at least a part of the risk information to the surrounding color of the sensitive information and displaying it, thereby anonymizing the risk information.

However, the present invention de-identifies personal information contained in images in various formats within a PDF file to fit the image format, and processes personal information within the image in at least one of mosaic processing, hiding processing, encryption processing, pseudonym processing, total processing, data deletion, data categorization, or masking, and the present invention crops or deletes only sensitive information within a medical image, when compared with the Korean Patent No. 10-2319492, which de-identifies personal information contained in images in various formats within a PDF file to fit the image format, and processes the personal information within the image in at least one of mosaic processing, hiding processing, encryption processing, pseudonym processing, total processing, data deletion, data categorization, or masking, and the Korean Patent No. 10-2585793, which performs cropping processing or deletion processing to extract and delete only sensitive information. There is a significant difference in the configuration.

The present invention was created to solve the above problems, and the purpose of the present invention is to provide a system and method capable of preventing exposure of personal information included in images in PDF files by de-identifying personal information included in images in various formats (i.e., TIFF, JPEG, PNG, and BMP) in accordance with the image format.

In particular, the present invention aims at providing a system and method capable of de-identifying personal information contained in images in various formats within a PDF file by extracting an image area contained in a PDF file, confirming the file format and coordinates, deleting personal information within the image area or overwriting it with another object, and then saving it using a file format generator, thereby performing de-identification, and inserting and updating the de-identified image into the original PDF file.

However, the technical tasks that this embodiment seeks to accomplish are not limited to the technical tasks described above, and other technical tasks may exist.

According to one embodiment of the present invention, a system for de-identifying images containing personal information in a document comprises: an image extraction unit for extracting at least one image containing personal information from a PDF file to be de-identified; a de-identification area determination unit for determining a de-identification area from the at least one image extracted; a de-identification processing unit for performing de-identification processing on the determined de-identification area; and a de-identification application unit for storing at least one image on which the de-identification processing has been performed through a file generator for each image, and updating the stored at least one image on which the de-identification processing has been performed in a corresponding image portion of the PDF file to be de-identified.

At this time, the personal information is characterized as information that can identify an individual, including name, address, phone number, resident registration number, email address, credit card number, date of birth, gender, occupation, income level, religion, political orientation, medical information, photograph, financial information, purchase information, and location information.

In addition, the image extraction unit is characterized by including an image file format verification unit that verifies which file format of the at least one or more extracted images is among TIFF, JPEG, PNG, and BMP; and a coordinate verification unit that verifies coordinate information of the at least one or more extracted images.

In addition, the anonymization application unit is characterized by storing the image performed in the anonymization processing unit through a file generator for each file format according to the file formats of TIFF, JPEG, PNG and BMP confirmed by the image file format confirmation unit.

In addition, the de-identification processing unit is characterized by performing de-identification processing of at least one de-identified area for each image based on coordinate information confirmed by the coordinate confirmation unit.

In addition, the image extraction unit inputs the PDF file to be de-identified into an artificial intelligence model for extracting personal information to extract at least one image containing personal information, and the artificial intelligence model for extracting personal information is characterized in that it is a deep learning model that has learned the characteristics of an image containing personal information.

In addition, the de-identification processing is characterized in that it is performed by removing, replacing, modifying, adding, or masking at least one image-specific de-identification area determined by the de-identification area determination unit.

In addition, a method for de-identifying an image containing personal information in a document according to an embodiment of the present invention is performed in a system for de-identifying an image containing personal information in a document, and is characterized by including an image extraction step of extracting at least one image containing personal information from a PDF file to be de-identified; a de-identification region determination step of determining a de-identification region from the at least one image extracted; a de-identification processing step of performing de-identification processing on the determined de-identification region; and a de-identification application step of storing the at least one image on which the de-identification processing was performed through a file generator for each image, and updating the stored at least one image on which the de-identification processing was performed in a corresponding image portion of the PDF file to be de-identified.

In addition, the image extraction step includes an image file format confirmation step of confirming which file format of the at least one or more extracted images is a TIFF, JPEG, PNG, and BMP; and a coordinate confirmation step of confirming coordinate information of the at least one or more extracted images; and the de-identification application step includes storing the image performed in the de-identification processing step through a file generator for each file format according to the file formats of TIFF, JPEG, PNG, and BMP confirmed in the image file format confirmation step, and the de-identification processing step is characterized in that it includes performing de-identification processing of the at least one or more image-specific de-identified regions based on the coordinate information confirmed in the coordinate confirmation step.

In addition, the image extraction step inputs the PDF file to be de-identified into an artificial intelligence model for extracting personal information to extract at least one image containing personal information, and the artificial intelligence model for extracting personal information is characterized in that it is a deep learning model that has learned the characteristics of an image containing personal information.

In addition, the de-identification processing is characterized in that it is performed by removing, replacing, modifying, adding, or masking at least one image-specific de-identification area determined in the de-identification area determination step.

As described above, according to the system and method for de-identifying images containing personal information in documents of the present invention, personal information contained in images in PDF files is de-identified according to the corresponding image formats of TIFF, JPEG, PNG and BMP, thereby fundamentally blocking easy removal of information hidden by inserting an opaque object over an existing image using an editing tool, thereby preventing exposure of personal information.

In addition, the present invention can significantly reduce the time and cost required to perform a task of de-identifying personal information contained in an image within a PDF file, and when modification or editing is required, only the necessary portion of the image can be selected and processed without having to repeat the task from the beginning, thereby improving work efficiency.

However, the effects of the present invention are not limited to the effects described above, and effects not mentioned can be clearly understood by a person having ordinary skill in the art to which the present invention pertains from this specification and the attached drawings.

FIG. 1 is a drawing schematically showing the entire configuration including an image de-identification system including personal information in a document according to one embodiment of the present invention.
FIG. 2 is a drawing showing in more detail the configuration of an image de-identification system according to one embodiment of the present invention.
FIG. 3 is a diagram showing the hardware structure of an image de-identification system according to one embodiment of the present invention.
FIG. 4 is a flowchart illustrating in detail the operation process of a method for de-identifying an image containing personal information in a document according to one embodiment of the present invention.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. However, the spirit of the present invention is not limited to the presented embodiments, and those skilled in the art who understand the spirit of the present invention will be able to easily propose other regressive inventions or other embodiments included within the scope of the spirit of the present invention by adding, changing, deleting, etc. other components within the scope of the same spirit, but this will also be considered to be included within the scope of the spirit of the present invention.

In addition, components having the same function within the same scope of the same idea that appear in the drawings of each embodiment are described using the same reference numerals.

FIG. 1 is a drawing schematically showing the entire configuration including an image de-identification system including personal information in a document according to one embodiment of the present invention.

As illustrated in FIG. 1, the present invention is configured to include an image de-identification system (100, hereinafter referred to as an image de-identification system) containing personal information in a document, a plurality of external servers (200), a database (300), a plurality of user terminals (400), etc.

The above image de-identification system (100) refers to a user-operated platform or server computer that performs de-identification processing to prevent personal information contained in images in documents used by the public, particularly PDF files, from being exposed.

The above image de-identification system (100) de-identifies personal information contained in images in multiple PDF files created based on various topics such as theses, reports, and articles held in multiple external servers (200) or an internal database (300) to match the image file format, thereby preventing personal information contained in images from being exposed when the documents are used by the public.

To this end, the image de-identification system (100) collects PDF files requiring de-identification processing from multiple external servers (200), its own database (300), and individual user terminals (400) through a network, or receives a request for de-identification processing from an external server operator or user.

Next, the image de-identification system (100) extracts an image portion containing personal information from a PDF file to be de-identified, de-identifies the extracted image portion containing personal information, and then updates the de-identified image back to the original PDF file to complete the de-identification process.

At this time, the images extracted from the PDF file are in image file formats such as TIFF, JPEG, PNG, and BMP.

For example, TIFF (Tagged Image File Format) image files can be used in fields where maintaining high quality is important, such as photography, medical imaging, and printing, because they support lossless compression to maintain image quality. JPEG (Joint Photographic Experts Group) can store files with small file sizes using lossy compression, so it is suitable for storing web images and general photos. PNG (Portable Network Graphics) can maintain image quality with lossless compression and supports transparency, making it suitable for web design. BMP (Windows Bitmap) has a simple structure and is supported by most operating systems and software, so it has high compatibility.

In the present invention, when personal information is included in an image in the TIFF, JPEG, PNG and BMP file formats within a PDF file, an anonymization operation such as deleting the image or overwriting it with another object is performed, and then an anonymized image is generated according to the image format using a file generator for each image.

And by de-identifying the personal information contained in the images in the PDF file by updating the de-identified images generated according to each image format back into the original PDF file (i.e., removing the original image portion and filling it with the newly generated de-identified image), it is possible to prevent specific personal information from being exposed when the general public uses the PDF file.

In this way, by saving an anonymized image with personal information removed using a dedicated file generator according to each image file format, rather than simply covering up the image object containing personal information, the image pixel values, feature vectors, and other information expressed as numbers are accurately calculated according to the image file format to accurately anonymize the image at the coordinates for which anonymization processing has been requested. Therefore, it is possible to fundamentally block the easy removal of information covered up by inserting an opaque object over an existing image object using an editing tool.

That is, the image de-identification technology applied to the present invention is not a method of simply covering an area requiring de-identification with an arbitrary object, but a method of generating and updating an de-identified image for each image file format by utilizing image encoding technology.

In addition, the image de-identification system (100) can perform a de-identification editing operation by selecting only some of the necessary images if the de-identification processing is not performed or modification is required after the de-identification processing of images containing personal information in a PDF file is performed. Through this, there is no need to repeat the de-identification operation from the beginning when modification or editing is required as in the existing method, so the time and cost required for the de-identification operation of images containing personal information can be greatly reduced, and work efficiency can be maximized.

The above external server (200) refers to a server operated by the government, local governments, companies, research institutes, hospitals, etc., and has multiple PDF files created on various topics such as theses, reports, and articles.

In addition, the external server (200) provides various PDF files to the image de-identification system (100) through a network to request de-identification of personal information included in images within documents, and returns and stores PDF files that have undergone de-identification processing.

Accordingly, administrators as well as the general public (i.e. user terminals) can now provide documents without exposing images containing personal information when providing PDF files upon request.

The above database (300) stores and manages various PDF files that have not been de-identified, and stores and updates various operation programs (e.g., image extraction program within PDF files, de-identification processing program, file generator for each image file format, etc.) used in the image de-identification system (100).

In addition, the database (300) stores and manages various PDF files for which image de-identification work performed by the image de-identification system (100) has been completed.

The above user terminal (400) is a communication terminal such as a smart phone or tablet PC used by the general public that utilizes various PDF files, and can search and use PDF files that have undergone image de-identification processing in the image de-identification system (100). At this time, the search and viewing of various PDF files can be done by connecting to the image de-identification system (100) or each external server (200).

In addition, the user terminal (400) can request an image de-identification task by providing a PDF file held by the user to the image de-identification system (100), and can receive a PDF file on which the image de-identification process has been completed.

FIG. 2 is a drawing showing in more detail the configuration of an image de-identification system according to one embodiment of the present invention.

As illustrated in FIG. 2, the image de-identification system (100) is configured to include a processing target file collection unit (110), an image extraction unit (120), a de-identification area determination unit (130), a de-identification processing unit (140), a de-identification application unit (150), a de-identification editing unit (160), etc.

The above processing target file collection unit (110) collects PDF files held by multiple external servers (200), PDF files held in the own database (300), PDF files held by individual users, etc. through a network. In other words, various PDF files that are targets of image de-identification processing are collected through various routes.

At this time, the processing target file collection unit (110) may receive a PDF file of the image de-identification processing target directly from an administrator or user of an external server and request image de-identification processing.

The above image extraction unit (120) extracts at least one image containing personal information from a PDF file of an image anonymization processing target collected through the processing target file collection unit (110).

At this time, the above personal information refers to information that can identify an individual as well as information that can be traced.

For example, personal information includes name, address, phone number, social security number, email address, credit card number, date of birth, gender, occupation, income level, religion, political orientation, medical information, photos, financial information, purchase information, and location information.

In addition, the image extraction unit (120) is configured to include an image file format verification unit (121) and a coordinate verification unit (122).

The above image file format verification unit (121) verifies which file format of an image containing personal information extracted from a PDF file, which is the subject of image de-identification processing, is among TIFF, JPEG, PNG, and BMP.

The above coordinate verification unit (122) verifies the coordinate information of an image containing personal information extracted from a PDF file that is the subject of image de-identification processing.

For example, the image extraction unit (120) can input a PDF file, which is the target of image de-identification processing, into a pre-prepared artificial intelligence model for extracting personal information and extract at least one image containing personal information.

At this time, the artificial intelligence model for extracting personal information is a deep learning model that has learned the characteristics of images containing various personal information.

Additionally, the features of the image containing the personal information may include not only the shape of a face, credit card, business card, etc., but also text with a pattern that matches the type of personal information.

In addition, the image extraction unit (120) may extract text from each image of a PDF document by performing OCR (Optical Character Recognition) processing on the PDF file, which is the target of image de-identification processing, and may also extract the text by comparing the extracted text with a pattern that matches a type of personal information such as a resident registration number, ID number, card number, name, date of birth, etc. prepared in advance.

The above-mentioned de-identified area determination unit (130) performs the function of determining a de-identified area in at least one image extracted by the image extraction unit (120).

For example, the de-identification area determination unit (130) can determine whether to de-identify the entire image extracted from the PDF file or to de-identify only a portion of the image extracted from the PDF file.

That is, if the entire image extracted from the PDF file consists of personal information, the entire image must be de-identified. However, if only a portion of the image is determined to be personal information, there may be cases where a portion of the image, not the entire image, must be de-identified. Therefore, the area to be de-identified is finally determined through the de-identification area determination unit (130).

The above-mentioned de-identification processing unit (140) performs de-identification processing on the de-identification area determined by the above-mentioned de-identification area determination unit (130).

That is, the de-identification processing unit (140) performs de-identification processing on the de-identified area of each image extracted from the PDF file based on the coordinate information confirmed by the coordinate confirmation unit (122) of the image extraction unit (120).

At this time, the above-described de-identification processing can be performed by removing, replacing, modifying, adding, or masking at least one image-specific de-identification area determined by the above-described de-identification area determination unit (130). Removing means deleting the image, and replacing, modifying, adding, and masking means overlaying an image containing personal information with another object (e.g., a black shape) or performing pixel processing.

The above-described de-identification application unit (150) stores each image on which de-identification processing has been performed in the above-described de-identification processing unit (140) through a file generator (not shown) for each image format file, and updates and stores each image on which de-identification processing has been performed in the corresponding image section of the PDF file that is the target of image de-identification processing.

That is, the anonymization application unit (150) stores the image performed in the anonymization processing unit (140) through a file generator for each file format according to the file formats of TIFF, JPEG, PNG and BMP confirmed by the image file format confirmation unit (120) of the image extraction unit (120).

For example, if the file format of the image to be processed as confirmed by the image file format confirmation unit (120) is a TIFF file, the image that has been de-identified is stored through a TIFF file generator, and if the file format of the image to be processed is a JPEG file, the image that has been de-identified is stored through a JPEG file generator.

The above-mentioned de-identification editing unit (160) provides a function that allows editing by selecting only the desired images among the images that have undergone de-identification processing through administrator operation, if editing is required for modification, deletion, or addition after at least one image that has undergone de-identification processing through the above-mentioned de-identification application unit (150) has been updated in the corresponding image portion of the PDF file that is the target of de-identification processing.

This is to improve work efficiency by allowing administrators to select and process only the necessary image portions without having to return the anonymized PDF file to its original state and repeat the work from the beginning.

FIG. 3 is a diagram showing the hardware structure of an image de-identification system according to one embodiment of the present invention.

As illustrated in FIG. 3, the hardware structure of the image de-identification system (100) includes a central processing unit (1000), memory (2000), user interface (3000), database interface (4000), network interface (5000), web server (6000), etc.

The above user interface (3000) provides an input and output interface to the user by using a graphical user interface (GUI).

The above database interface (4000) provides an interface between the database and the hardware structure. The above network interface (5000) provides a network connection between devices held by users.

The above web server (6000) provides a means for users to access the hardware structure through a network. Most users can access the image de-identification system (100) remotely by accessing the web server.

Each step of the above-described configuration or method may be implemented as a computer-readable code on a computer-readable recording medium or transmitted via a transmission medium. The computer-readable recording medium is a data storage device capable of storing data that can be read by a computer system.

Examples of computer-readable recording media include, but are not limited to, databases, ROMs, RAMs, CD-ROMs, DVDs, magnetic tapes, floppy disks, and optical data storage devices. Transmission media may include carrier waves transmitted over the Internet or various types of communication channels. Additionally, the computer-readable recording media may be distributed through a network-coupled computer system so that the computer-readable code is stored and executed in a distributed manner.

In addition, at least one or more components applied to the present invention may include or be implemented by a processor such as a central processing unit (CPU), a microprocessor, etc., which performs each function, and two or more of the components may be combined into a single component that may perform all operations or functions of the two or more combined components. In addition, a part of at least one or more components applied to the present invention may be performed by another component among these components. In addition, communication between the components may be performed through a bus (not shown).

Next, an embodiment of a method for de-identifying an image containing personal information in a document according to the present invention configured as described above will be described in detail with reference to FIG. 4. At this time, the order of each step according to the method of the present invention may be changed depending on the usage environment or a person skilled in the art.

FIG. 4 is a flowchart illustrating in detail the operation process of a method for de-identifying an image containing personal information in a document according to one embodiment of the present invention.

As illustrated in FIG. 4, the image de-identification system (100) collects PDF files on various topics that are subject to image de-identification processing from multiple external servers (200), its own database (300), and individual users through a network (S100).

At this time, the image de-identification system (100) may, in addition to collecting PDF files, receive various PDF files and request image de-identification processing.

The image de-identification system (100) that collects or requests various PDF files to be subject to image de-identification processing through the above step S100 extracts at least one image containing personal information from each PDF file and checks which file format each image is in among TIFF, JPEG, PNG, and BMP, and what the coordinate information for each image is (S200).

As explained above, at this time, the extraction of images containing personal information is done by inputting a specific PDF file that is the subject of anonymization processing into an artificial intelligence model for extracting personal information that has learned the characteristics of images containing personal information.

Next, the image de-identification system (100) determines a de-identified area in at least one image extracted in step S200 (S300). That is, all or part of an image containing personal information extracted from a PDF file is determined as a de-identified area.

In addition, the image de-identification system (100) performs de-identification processing on the de-identified area determined through the S300 step (S400). That is, based on the coordinate information confirmed in the S200 step, the de-identification processing is performed on the de-identified area determined in the S300 step by removing, replacing, modifying, adding, or masking the area.

In addition, the image de-identification system (100) determines whether the de-identification processing of the image containing personal information extracted from the PDF file through the step S400 is completed (S500), and if the de-identification processing is completed, the image on which the de-identification processing was performed is stored through the corresponding image-specific file generator (S600).

Next, the image de-identification system (100) updates and stores each de-identified image saved through the image-specific file generator in the corresponding image portion of the PDF file to be de-identified (S700).

In addition, the image de-identification system (100) updates and saves the de-identified image in the original PDF file through the S700 step, and then, if editing for modification, deletion, and addition is required, only the desired image among the images that have undergone de-identification processing is selected and edited by referring to the administrator's operation (S800).

In this way, the present invention can prevent personal information exposure by fundamentally blocking the easy removal of information hidden by inserting an opaque object over an existing image using an editing tool, and not only reduces the time and cost required to perform a task of anonymizing personal information included in an image within a PDF file, but also improves work efficiency by selecting and processing only the necessary image portion without having to repeat the task from the beginning when modification or editing is required.

In order to more clearly express the technical idea of the present invention, the attached drawings briefly express or omit components that are not related to or have little to do with the technical idea of the present invention.

Although the configuration and features of the present invention have been described above based on embodiments according to the present invention, the present invention is not limited thereto, and it will be apparent to those skilled in the art that various changes or modifications can be made within the spirit and scope of the present invention, and therefore it is made clear that such changes or modifications fall within the scope of the appended patent claims.

100: Image de-identification system containing personal information in documents
110: Processing target file collection section
120 : Image Extraction Section
121: Image file format verification section
122: Coordinate confirmation section
130: Anonymization area determination section
140: Anonymization Processing Unit
150: Anonymization application area
160: Anonymization Editorial Department
200 : External server
300 : Database
400: User Terminal

Claims (11)

An image extraction unit that inputs a PDF file to be de-identified into an artificial intelligence model for extracting personal information, which is a deep learning model that has learned the characteristics of images containing personal information, and extracts at least one image containing personal information, and at the same time verifies coordinate information of at least one image extracted, and verifies which file format it is among TIFF, JPEG, PNG, and BMP;
An anonymized region determining unit for determining an anonymized region from at least one image extracted above;
A de-identification processing unit that performs de-identification processing on an unidentified region by removing, replacing, modifying, adding, or masking at least one image-specific unidentified region based on coordinate information confirmed by the image extraction unit;
A de-identification application unit that saves at least one image that has undergone the de-identification process in the file formats of TIFF, JPEG, PNG and BMP confirmed by the image extraction unit through a file generator for each file format, and updates the at least one image that has undergone the de-identification process in the corresponding image portion of the PDF file that is the target of the de-identification process; and
A system for de-identifying images containing personal information in a document, characterized in that it includes a de-identification editing unit; wherein, if anonymization processing of an image is not performed and thus additions or modifications and deletions are required after at least one image on which the de-identification processing has been performed is updated in the corresponding image portion of the PDF file to be de-identified, only some of the images are selected and edited for addition, modification, and deletion, thereby increasing the time and speed of the de-identification task by selecting and editing only the necessary image portions without having to return the PDF file to its original state and repeat the task from the beginning.
In claim 1,
The above personal information is,
An image de-identification system containing personal information in a document, characterized by information that can identify an individual, including name, address, phone number, resident registration number, email address, credit card number, date of birth, gender, occupation, income level, religion, political orientation, medical information, photograph, financial information, purchase information, and location information.
delete delete delete delete delete As performed in an image de-identification system containing personal information in a document as described in claim 1,
An image extraction step of inputting a PDF file to be de-identified into an artificial intelligence model for extracting personal information, which is a deep learning model that has learned the characteristics of images containing personal information, to extract at least one image containing personal information, and at the same time confirming the coordinate information of the at least one image extracted, and confirming which file format it is among TIFF, JPEG, PNG, and BMP;
A de-identified region determination step of determining a de-identified region from at least one image extracted above;
A de-identification processing step for performing de-identification processing on an unidentified region by removing, replacing, modifying, adding, or masking at least one image-specific unidentified region based on coordinate information confirmed in the image extraction step;
A de-identification application step of saving at least one image that has undergone the de-identification process in the file formats of TIFF, JPEG, PNG and BMP confirmed in the image extraction step through a file generator for each file format, and updating the at least one image that has undergone the de-identification process in the corresponding image portion of the PDF file that is the target of the de-identification process; and
A method for de-identifying images containing personal information in a document, characterized in that it includes a de-identification editing step, wherein, if anonymization processing of an image is not performed and thus additions or modifications and deletions are required after at least one image on which the de-identification processing has been performed is updated in the corresponding image portion of the de-identification processing target PDF file, only some of the images are selected and edited for addition, modification, and deletion, thereby increasing the time required and speed of the de-identification task by selecting and editing only the necessary image portions without having to return the de-identification processing target PDF file to its original state and repeating the task from the beginning. delete delete delete