Method for de-identifying identification information and apparatus for performing same

Abstract

Disclosed are a method for de-identifying identification information and an apparatus for performing same. The method according to an embodiment may comprise the operations of: detecting identification information enabling identification of an object included in original media; classifying the identification information into first identification information and second identification information; generating virtual identification information to replace the identification information, by respectively inputting the first identification information and the second identification information to a model; and on the basis of the virtual identification information, generating a result in which the identification information has been de-identified. The first identification information may be direct information enabling recognition of the object, and the second identification information may be indirect information enabling inference of the object.

Description

Method for de-identifying identification information and apparatus for performing the same

[0001] The following disclosure relates to a method for de-identifying identification information and an apparatus for performing the same.

[0002] If content containing personal information (e.g., images and videos) is used as is, privacy infringement issues may arise. Existing de-identification methods, which simply delete or blur personal information, can significantly reduce the utility value of the data. Furthermore, existing methods have limitations in that they cannot achieve complete de-identification because they only process direct personal information (e.g., faces, names, and license plates). Therefore, there is a need to develop technology that performs de-identification on information capable of inferring specific individuals, in addition to direct personal information, while maintaining the utility value of the data.

[0003] The background technology described above is possessed or acquired by the inventor in the process of deriving the content of the disclosure of the present application, and cannot necessarily be considered as prior art disclosed to the general public prior to the filing of this application.

[0004] One embodiment may provide a technology for processing identification information contained in the original media.

[0005] One embodiment may provide a technology for generating virtual identification information to replace indirect identification information that can be used to infer an individual.

[0006] One embodiment can de-identify identification information while maintaining the utility value of the original media by using context information.

[0007] However, technical challenges are not limited to the technical challenges described above, and other technical challenges may exist.

[0008] A de-identification method according to one embodiment may include detecting identification information capable of identifying an object included in an original media, classifying the identification information into a first identification information and a second identification information, inputting the first identification information and the second identification information respectively into a model to generate virtual identification information to replace the identification information, and generating a result in which the identification information is de-identified based on the virtual identification information. The first identification information may be direct information capable of recognizing the object, and the second identification information may be indirect information capable of inferring the object.

[0009] According to one embodiment, the operation of generating the virtual identification information may include the operation of generating the first virtual identification information by inputting the first identification information into the model and the operation of generating the second virtual identification information by inputting the second identification information into the model.

[0010] According to one embodiment, the operation of generating the second virtual identification information may include the operation of extracting context information from the second identification information and the operation of generating the second virtual identification information while maintaining the context information.

[0011] According to one embodiment, the operation of detecting the object may include a labeling operation of the object based on the type of the object.

[0012] According to one embodiment, the first identification information may include information about a face when the object is a person, and information about a text plate when the object is a vehicle.

[0013] According to one embodiment, the second identification information may include the person's hairstyle, clothes worn by the person, and accessories when the object is a person, and may include the type of vehicle and the color of the vehicle when the object is a vehicle.

[0014] According to one embodiment, the second identification information may further include text that can infer the location where the original media was filmed.

[0015] According to one embodiment, the text may include text included in at least one of a sign and a signpost included in the original media.

[0016] According to one embodiment, a computer-readable recording medium storing one or more computer programs may include instructions for performing the method in a processor.

[0017] An apparatus according to one embodiment may include at least one processor and a memory comprising instructions. Based on the instructions being executed individually or collectively by the at least one processor, the apparatus may detect identification information capable of identifying an object contained in the original media, classify the identification information into a first identification information and a second identification information, input the first identification information and the second identification information respectively into a model to generate virtual identification information to replace the identification information, and generate a result in which the identification information is de-identified based on the virtual identification information. The first identification information may be direct information capable of recognizing the object, and the second identification information may be indirect information capable of inferring the object.

[0018] According to one embodiment, based on the instructions being executed individually or collectively by at least one processor, the device may input the first identification information into the model to generate the first virtual identification information, and input the second identification information into the model to generate the second virtual identification information.

[0019] According to one embodiment, based on the instructions being executed individually or collectively by the at least one processor, the device may be able to extract context information from the second identification information and generate the second virtual identification information while maintaining the context information.

[0020] According to one embodiment, the instructions may be executed individually or collectively by the at least one processor, thereby enabling the device to label the object based on the type of the object.

[0021] According to one embodiment, the first identification information may include information about a face when the object is a person, and information about a text plate when the object is a vehicle.

[0022] According to one embodiment, the second identification information may include the person's hairstyle, clothes worn by the person, and accessories when the object is a person, and may include the type of vehicle and the color of the vehicle when the object is a vehicle.

[0023] According to one embodiment, the second identification information may further include text that can infer the location where the original media was filmed.

[0024] According to one embodiment, the text may include text included in at least one of a sign and a signpost included in the original media.

[0025] In relation to the description of the drawings, the same or similar reference numerals may be used for identical or similar components.

[0026] FIG. 1 is a schematic block diagram of a de-identification device according to one embodiment.

[0027] FIG. 2 is a diagram illustrating a model for classifying identification information according to one embodiment.

[0028] FIG. 3 is a diagram illustrating a de-identification model according to one embodiment.

[0029] FIGS. 4a and 4b are drawings for explaining the operation of generating virtual identification information according to one embodiment.

[0030] FIG. 5 is a diagram illustrating the operation of generating a result in which identification information is de-identified according to one embodiment.

[0031] FIG. 6 is a flowchart illustrating a de-identification method according to one embodiment.

[0032] FIG. 7 is a schematic block diagram of an electronic device according to one embodiment.

[0033] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, actual implementations are not limited to the specific embodiments disclosed, and the scope of this specification includes modifications, equivalents, or substitutions included in the technical concept described by the embodiments.

[0034] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0035] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or joined to that other component, or that there may be other components in between.

[0036] Singular expressions include plural expressions unless the context clearly indicates otherwise. In this document, phrases such as “A or B,” “at least one of A and B,” “at least one of A or B,” “A, B or C,” “at least one of A, B and C,” and “at least one of A, B, or C” may each include any one of the items listed together with the corresponding phrase, or all possible combinations thereof. In this specification, terms such as “comprising” or “having” are intended to designate the existence of the described feature, number, step, action, component, part, or combination thereof, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0037] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this specification.

[0038] As used herein, the term "module" may include a unit implemented in hardware, software, or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit. A module may be a component formed integrally, or a minimum unit of said component or a part thereof that performs one or more functions. For example, according to one embodiment, a module may be implemented in the form of an application-specific integrated circuit (ASIC).

[0039] As used in this document, the term "part" refers to a software or hardware component, such as an FPGA or ASIC, that performs certain roles. However, "part" is not limited to software or hardware. "Part" may be configured to reside in an addressable storage medium or configured to operate one or more processors. For example, "part" may include components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functions provided within the components and "parts" may be combined into a smaller number of components and "parts" or further separated into additional components and "parts." Furthermore, components and "parts" may be implemented to operate one or more CPUs within a device or secure multimedia card. Additionally, '~part' may include one or more processors.

[0040] Hereinafter, embodiments will be described in detail with reference to the attached drawings. In the description with reference to the attached drawings, identical components are given the same reference numeral regardless of the drawing number, and redundant descriptions thereof will be omitted.

[0041]

[0042] FIG. 1 is a schematic block diagram of a de-identification device according to one embodiment.

[0043] Referring to FIG. 1, according to one embodiment, a de-identification device (100) can process original media to produce a result in which identification information contained in the original media is de-identified. The de-identification device (100) can obtain (e.g., receive) the original media from another external device and / or server. The original media may contain visual data such as images or videos. For example, the original media may contain various types of visual data collected (e.g., captured) through various devices such as autonomous vehicles, drones, CCTVs, and mobile devices. The de-identification device (100) can preprocess the original media. For example, the de-identification device (100) may perform preprocessing such as standardization of the data format of the original media and / or resolution adjustment.

[0044] According to one embodiment, the de-identification device (100) may include a classification model (120), a de-identification model (140), and a synthesis unit (160). The classification model (120), the de-identification model (140), and the synthesis unit (160) may be software modules implemented on a processor of the de-identification device (100) (e.g., the processor (730) of FIG. 7). The classification model (120), the de-identification model (140), and the synthesis unit (160) are illustrated as examples for explaining the de-identification device (100), and may include various variations as long as the operations of the de-identification device (100) described in this disclosure can be implemented. For example, two or more components may be combined, or one or more components may be added or omitted.

[0045] According to one embodiment, the classification model (120) and the de-identification model (140) may be located inside and / or outside the de-identification device (100). For example, the classification model (120) and the de-identification model (140) may be implemented on another device and / or server outside the de-identification device (100). The classification model (120) and the de-identification model (140) are artificial intelligence (AI) models and may be implemented using a neural network. For example, the classification model (120) and the de-identification model (140) may be artificial intelligence models implemented using a generative adversarial network (GAN), but are not limited thereto.예를 들어, 뉴럴 네트워크는 심층 뉴럴 네트워크 (deep neural network), CNN(convolutional neural network), RNN(recurrent neural network), 퍼셉트론(perceptron), 다층 퍼셉트론(multilayer perceptron), FF(feed forward), RBF(radial basis network), DFF(deep feed forward), LSTM(long short term memory), GRU(gated recurrent unit), AE(auto encoder), VAE(variational auto encoder), DAE(denoising auto encoder), SAE(sparse auto encoder), MC(markov chain), HN(hopfield network), BM(boltzmann machine), RBM(restricted boltzmann machine), DBN(deep belief network), DCN(deep convolutional network), DN(deconvolutional network), DCIGN(deep convolutional inverse graphics network), LSM(liquid state machine), ELM(extreme learning machine), ESN(echo state network), DRN(deep residual network), DNC(differentiable neural computer), NTM(neural turning machine), CN(capsule network), KN(kohonen network) 및 AN(attention network)를 포함할 수 있다.

[0046] According to one embodiment, the de-identification device (100) can detect objects contained in the original media by using an artificial intelligence-based object detection algorithm. For example, the de-identification device (100) can detect objects subject to de-identification, such as people, vehicles, signs, and signposts contained in the original media, through an object detection model (e.g., the object detection model (240) of FIG. 2) that uses a deep-learning-based object detection algorithm. In response to the fact that no objects subject to de-identification are detected from the original media, the de-identification device (100) may not perform a de-identification method and may maintain the original media as is. In response to the detection of an object, the de-identification device (100) may store the location information and confidence score of the object. The location information of the object may be expressed as the coordinates of the point where the object is located within the original media (e.g., image, video). The de-identification device (100) can label the object based on the type of the detected object. The de-identification device (100) can detect identification information that can identify the object contained in the original media. The identification information may be unique information specific to the object that allows a particular object to be distinguished from other objects. For example, if the object is a person, the identification information may include information about items such as the person's name, face, hairstyle, clothes worn, and accessories that can be used to infer the person. For example, if the object is a signboard and / or sign, the identification information may be text containing information about the business name (e.g., AAA Gukbap), region, and place included in the signboard and / or sign.

[0047] According to one embodiment, a de-identification device (100) can classify identification information into a first identification information and a second identification information. The de-identification device (100) can classify identification information into the first identification information and the second identification information using a classification model (120). The first identification information may be direct information that can recognize an object contained in the original media. If the object is a person, the first identification information may include information about the face. If the object is a vehicle, the first identification information may include information about the text plate. The second identification information may be indirect information that can infer an object contained in the original media. If the object is a person, the second identification information may include items such as the person's hairstyle, clothes worn by the person, and accessories, but is not limited thereto. For example, if the object is a person, the second identification information may include various information that can be used to infer and identify that person. If the object is a vehicle, the second identification information may include the type of vehicle and the color of the vehicle. The second identification information may include text that can be used to infer the location where the original media was filmed. The text may include text contained in at least one of the signboards and signs included in the original media. For example, since the location and / or region where the original media was filmed can be inferred from the text "AAA Gukbap" included in the original media, the text "AAA Gukbap" may be included in the second identification information. Since the location where the original media was filmed and / or the location of an individual can be identified through the text contained in the signboards and signs, the second identification information may include text contained in at least one of the signboards and signs included in the original media.

[0048] According to one embodiment, a de-identification device (100) may input a first identification information and a second identification information respectively into a model (e.g., a de-identification model (140)) to generate virtual identification information to replace the identification information. The virtual identification information may include a first virtual identification information and a second virtual identification information. The de-identification device (100) may input the first identification information into the de-identification model (140) to generate the first virtual identification information. The de-identification device (100) may input the second identification information into the de-identification model (140) to generate the second virtual identification information. Based on the virtual identification information, the de-identification device (100) may generate a result in which the identification information is de-identified. The de-identification device (100) can input a first virtual identification information and a second virtual identification information into a synthesis unit (160) to generate a result in which the identification information is de-identified. For example, the synthesis unit (160) can generate a de-identified result by replacing the first identification information with the first virtual identification information and replacing the second identification information with the second virtual identification information. The de-identification device (100) can generate a result in which the overall context and utility value of the original media are maintained while minimizing the possibility of identifying an individual and / or location (e.g., the place where the original media was filmed) by de-identifying all information that can directly identify an object in the original media as well as information that can indirectly infer an object. The result generated by the de-identification device (100) can be utilized in various fields such as the training of artificial intelligence models and statistical analysis.

[0049]

[0050] FIG. 2 is a diagram illustrating a model for classifying identification information according to one embodiment.

[0051] Referring to FIG. 2, according to one embodiment, a classification model (220) (e.g., the classification model (120) of FIG. 1) may include an object detection model (240) and an identification information classification model (260). The object detection model (240) may detect objects contained in the original media using an object detection algorithm. The objects may include objects subject to de-identification. For example, the object detection model (240) may detect objects such as people, vehicles, and signs and billboards containing text, which are objects subject to de-identification from the original media. The object detection model (240) may detect signs and billboards containing text from the original media through optical character recognition (OCR) technology. The text contained in the signs and billboards may be text that can infer the location where the original media was filmed. For example, the text may be text containing information regarding a business name (e.g., AAA Gukbap), region, or place contained in the signs and / or billboards.

[0052] According to one embodiment, a detected object (e.g., an object such as a person, vehicle, signboard, or sign) may be transmitted to an identification information classification model (260). The identification information classification model (260) may label the object based on the type of the detected object. For example, in response to the detected object being a person, the identification information classification model (260) may label the object as a person and may also label the identification information capable of identifying the person, such as face, hairstyle, clothes worn, and accessories. For example, in response to the detected object being a vehicle, the identification information classification model (260) may use a vehicle recognition algorithm to recognize the license plate, type, and color of the vehicle and label each piece of information about the vehicle. Based on the labeling results for the identification information, the identification information capable of identifying the object may be classified into first identification information and second identification information. For example, among the identification information capable of identifying a person, which is a detected object, the identification information can classify information about the face, which can directly identify the person, as first identification information, and indirect information that can infer the person, such as hairstyle, clothes worn by the person, and accessories, as second identification information. For example, among the identification information capable of identifying a vehicle, which is a detected object, the identification information can classify information about the license plate, which can directly identify the vehicle, as first identification information, and indirect information that can infer the vehicle, such as the type of vehicle and the color of the vehicle, as second identification information. If the object is a vehicle, the identification information classification model (260) may include text that can infer the location where the original media was filmed as second identification information. Text included in at least one of the signboards and signs included in the original media can be classified as second identification information because it is information that can indirectly identify a person and / or location.

[0053]

[0054] FIG. 3 is a diagram illustrating a de-identification model according to one embodiment.

[0055] Referring to FIG. 3, according to one embodiment, a de-identification model (340) (e.g., the de-identification model (140) of FIG. 1) can generate virtual identification information to replace identification information based on identification information. The de-identification model (340) can receive first identification information and generate first virtual identification information. The de-identification model (340) can receive second identification information and generate second virtual identification information. The de-identification model (340) is an artificial intelligence model (e.g., a generative model) based on a neural network (e.g., a neural network such as a GAN) and can generate virtual identification information (e.g., first virtual identification information, second virtual identification information) to replace identification information. The virtual identification information may be generated to replace identification information. For example, the virtual identification information may maintain the class of identification information so that what it is can be distinguished, but the identification information may be de-identified so that it is impossible to know who or where it is by changing the instance. The de-identification device (100) can effectively protect identification information while preserving the statistical characteristics and learning performance of the original media by generating virtual identification information. The operation of generating the first virtual identification information and the second virtual identification information will be explained in detail with reference to FIG. 4a and FIG. 4b, respectively.

[0056]

[0057] FIGS. 4a and 4b are drawings for explaining the operation of generating virtual identification information according to one embodiment.

[0058] Referring to FIGS. 4a and 4b, according to one embodiment, a de-identification model (440) (e.g., the de-identification model (140) of FIG. 1) may receive first identification information to generate first virtual identification information and receive second identification information to generate second virtual identification information. The de-identification model (440) may include a context extraction unit (443) and a generation unit (445). The context extraction unit (443) may extract context information from the first identification information and / or the second identification information. The context information may be additional information such as style, size, font, and background color. The generation unit (445) may be implemented as an artificial intelligence-based generative model. For example, the generation unit (445) may be implemented as a neural network-based generative model such as a GAN.

[0059] According to one embodiment, the de-identification model (440) can receive first identification information and generate first virtual identification information. For example, the generation unit (445) of the de-identification model (440) can generate a virtual face and a virtual license plate that can replace a person's face and a vehicle's license plate detected from the original media, while being unable to identify a specific object. The generation unit (445) can generate a virtual license plate while maintaining context information extracted from the vehicle's license plate (e.g., context information such as the shape, color, and font of the license plate). For example, the virtual license plate, which is the generated first virtual identification information, may be a license plate included in the original media in which only the vehicle number has been changed from '1234' to '5678'.

[0060] According to one embodiment, the de-identification model (440) can receive second identification information and generate second virtual identification information. For example, the generating unit (445) of the de-identification model (440) can generate a virtual hairstyle, virtual clothing, and virtual accessories that can replace a person's hairstyle, clothing worn by a person, and accessories detected from the original media, while not identifying a specific object. For example, the generating unit (445) of the de-identification model (440) can generate a virtual vehicle in which the type of vehicle and the color of the vehicle detected from the original media have been changed. For example, the generating unit (445) of the de-identification model (440) can generate a virtual signboard and / or signboard in which the text included in the original media has been changed. The de-identification model (440) can extract context information of the second identification information. For example, the de-identification model (440) can extract context information through the context extraction unit (443) to generate second virtual identification information similar to the second identification information. The context information extracted from the second identification information can be transmitted to the generation unit (445). The de-identification model (440) can generate second virtual identification information while maintaining the context information. For example, the generation unit (445) of the de-identification model (440) can generate virtual clothing that is similar in style (e.g., sleeveless top, pants) to the clothing worn by a person in the original media, but differs through changes in pattern and / or color. For example, the de-identification model (440) can generate virtual hairstyles and virtual accessories that can replace the person's hairstyle and accessories (e.g., suitcase) in the original media. For example, the de-identification model (440) can generate a virtual trade name “BBB Gukbap” that can replace the trade name “AAA Gukbap” included in the original media, and in this case, “BBB Gukbap” may not be able to identify a specific region and / or place.The generated first virtual identification information and second virtual identification information can be transmitted to a synthesis unit (e.g., the synthesis unit (160) of FIG. 1).

[0061]

[0062] FIG. 5 is a diagram illustrating the operation of generating a result in which identification information is de-identified according to one embodiment.

[0063] Referring to FIG. 5, according to one embodiment, a synthesis unit (560) (e.g., the synthesis unit (160) of FIG. 1) can generate a result in which identification information is de-identified based on virtual identification information (e.g., first virtual identification information, second virtual identification information). The synthesis unit (560) can generate a result in which the generated virtual identification information is synthesized into the original media using an artificial intelligence-based generative model. For example, the synthesis unit (560) can be implemented as a neural network-based generative model such as a GAN. The synthesis unit (560) can generate a result in which the virtual identification information is replaced with an object in the original media using location information of an object within the original media. For example, the result in which identification information is de-identified may be one in which the identification information is de-identified by changing the person's face, hairstyle, clothes and accessories, vehicle license plate and color included in the original media, and changing the text included in the original media.

[0064]

[0065] FIG. 6 is a flowchart illustrating a de-identification method according to one embodiment.

[0066] Referring to FIG. 6, according to one embodiment, operations 610 to 670 may be operations performed by the de-identification device (100) of FIG. 1 described with reference to FIG. 1 to FIG. 5.

[0067] According to one embodiment, operations 610 to 670 may be understood to be performed in a processor (e.g., processor (730) of FIG. 7) of a de-identification device (100) (e.g., electronic device (700) of FIG. 7) described with reference to FIG. 1 to 5.

[0068] In operation 610, the de-identification device (100) can detect identification information that can identify an object contained in the original media.

[0069] In operation 630, the de-identification device (100) may classify identification information into first identification information and second identification information. The first identification information is direct information that can recognize an object, and the second identification information may be indirect information that can infer an object.

[0070] In operation 650, the de-identification device (100) may input the first identification information and the second identification information into the model, respectively, to generate virtual identification information to replace the identification information. The virtual identification information may include the first virtual identification information and the second virtual identification information.

[0071] In operation 670, the de-identification device (100) can generate a result in which the identification information is de-identified based on virtual identification information. The de-identification device (100) can generate a result in which the identification information is protected while maintaining the statistical characteristics and utility value of the original media.

[0072] Operations 610 through 670 may be performed sequentially, but are not limited thereto. For example, two or more operations may be performed in parallel.

[0073]

[0074] FIG. 7 is a schematic block diagram of an electronic device according to one embodiment.

[0075] Referring to FIG. 7, according to one embodiment, an electronic device (700) (e.g., a de-identification device (100) of FIG. 1) may include a memory (710) and a processor (730).

[0076] The memory (710) can store instructions (or programs) executable by the processor (730). For example, the instructions may include instructions for executing the operation of the processor (730) and / or the operation of each component of the processor (730).

[0077] The memory (710) may include one or more computer-readable storage media. The memory (710) may include non-volatile storage devices (e.g., magnetic hard disc, optical disc, floppy disc, flash memory, EPROM (electrically programmable memories), EEPROM (electrically erasable and programmable)).

[0078] The memory (710) may be a non-transitory medium. The term "non-transitory" may indicate that the storage medium is not implemented by a carrier wave or a propagated signal. However, the term "non-transitory" should not be interpreted as meaning that the memory (710) is immobile.

[0079] The processor (730) can process data stored in memory (710). The processor (730) can execute computer-readable code (e.g., software) stored in memory (710) and instructions triggered by the processor (730).

[0080] The processor (730) may be a data processing device implemented in hardware having a circuit having a physical structure for executing desired operations. For example, the desired operations may include code or instructions included in a program.

[0081] For example, a data processing device implemented in hardware may include a microprocessor, a central processing unit, a processor core, a multi-core processor, a multiprocessor, an Application-Specific Integrated Circuit (ASIC), and a Field Programmable Gate Array (FPGA).

[0082] The processor (730) can cause the electronic device (700) to perform one or more operations by executing code and / or instructions stored in memory (710). The operations performed by the electronic device (700) may be substantially the same as the operations performed by the de-identification device (100) described with reference to FIGS. 1 through 7. Such redundant descriptions are omitted.

[0083]

[0084] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using a general-purpose computer or a special-purpose computer, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. In addition, other processing configurations, such as parallel processors, are also possible.

[0085] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or instruct the processing unit independently or collectively. Software and / or data may be stored on any type of machine, component, physical device, virtual equipment, computer storage medium, or device so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on computer-readable recording media.

[0086] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may store program instructions, data files, data structures, etc., either individually or in combination, and the program instructions recorded on the medium may be those specifically designed and configured for the embodiment or those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

[0087] The hardware device described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.

[0088] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based thereon. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

[0089] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below.

Claims

1. Detecting identification information capable of identifying objects contained in the original media; The operation of classifying the above identification information into a first identification information and a second identification information; The operation of inputting the first identification information and the second identification information into the model respectively to generate virtual identification information to replace the identification information; and Based on the above-mentioned virtual identification information, the operation of generating a de-identified result of the identification information. Includes, The above first identification information is, It is direct information that can recognize the above object, and The above second identification information is, A de-identification method that is indirect information capable of inferring the above object.

2. In Paragraph 1, The operation of generating the above-mentioned virtual identification information is, The operation of inputting the above-mentioned first identification information into the above-mentioned model to generate first virtual identification information; and The operation of inputting the above second identification information into the model to generate second virtual identification information A de-identification method including 3. In Paragraph 2, The operation of generating the above-mentioned second virtual identification information is, The operation of extracting context information from the second identification information; and The operation of generating the second virtual identification information while maintaining the above context information A de-identification method including 4. In Paragraph 1, The operation of detecting the above object is, Labeling the object based on the type of the object A de-identification method including 5. In Paragraph 1, The above first identification information is, If the above object is a person, it includes information about the face, and A de-identification method comprising information about a text plate when the object is a vehicle.

6. In Paragraph 1, The above second identification information is, If the above object is a person, it includes the person's hairstyle, clothes worn by the person, and accessories, and A de-identification method comprising the type of vehicle and the color of the vehicle, where the object is a vehicle.

7. In Paragraph 1, The above second identification information is, Text that allows one to infer the location where the above original media was filmed A de-identification method including further 8. In Paragraph 7, The above text is, A de-identification method comprising text included in at least one of a sign and a signpost included in the original media.

9. A computer program stored on a computer-readable recording medium in combination with hardware to execute the method of any one of claims 1 through 8.

10. In the device, At least one processor; and memory that stores instructions Includes, Based on the above instructions being executed individually or collectively by the at least one processor, the device, Detect identification information capable of identifying objects contained in the original media, and The above identification information is classified into first identification information and second identification information, and The above first identification information and the above second identification information are respectively input into the model to generate virtual identification information to replace the identification information, and Based on the above-mentioned virtual identification information, the identification information is de-identified to generate a result, and The above first identification information is, It is direct information that can recognize the above object, and The above second identification information is, A device that is indirect information capable of inferring the above object.

11. In Paragraph 10, Based on the above instructions being executed individually or collectively by the at least one processor, the device, The above first identification information is input into the model to generate first virtual identification information, and A device that inputs the above-mentioned second identification information into the above-mentioned model to generate second virtual identification information.

12. In Paragraph 11, Based on the above instructions being executed individually or collectively by the at least one processor, the device, Context information is extracted from the above second identification information, and A device that generates the second virtual identification information while maintaining the above context information.

13. In Paragraph 10, Based on the above instructions being executed individually or collectively by the at least one processor, the device, A device for labeling the object based on the type of the object.

14. In Paragraph 10, The above first identification information is, If the above object is a person, it includes information about the face, and A device that includes information about a text plate, where the object is a vehicle.

15. In Paragraph 10, The above second identification information is, If the above object is a person, it includes the person's hairstyle, clothes worn by the person, and accessories, and A device comprising, where the object is a vehicle, the type of the vehicle and the color of the vehicle.

16. In Paragraph 10, The above second identification information is, Text that allows one to infer the location where the above original media was filmed A device that further includes 17. In Paragraph 16, The above text is, A device comprising text included in at least one of a sign and a signpost included in the original media.

Images