KR102340984B1 - An apparatus for wound management using machine learning and operations thereof - Google Patents

Abstract

The present disclosure relates to an apparatus for wound management and a method of operating the apparatus for wound management, and the present disclosure relates to an operating method of the apparatus for wound management, comprising: acquiring image data of a plurality of existing wounds; Obtaining treatment information of an existing wound corresponding to the image data, acquiring a machine learning model by learning the correlation between image data of a plurality of existing wounds and treatment information of existing wounds with a CNN (Convolution Neural Network); Acquiring the patient's wound image data, applying a machine learning model to the current patient's wound image data to determine wound treatment prediction information, and outputting wound treatment prediction information, the treatment information of the existing wound contains all of the wound depth information, the wound healing period information, the information on whether surgery is performed, and the information on whether or not there is a scar, and the wound treatment prediction information includes the expected information on the presence or absence of surgery and the expected information on the occurrence of scars. characterized in that

Description

An apparatus for wound management using machine learning and an operating method of the apparatus for wound management {An apparatus for wound management using machine learning and operations thereof}

The present disclosure relates to an apparatus for wound management and a method of operation of the apparatus for wound management.

Wounds are classified according to the cause into cleavage (切創), puncture (刺創), split spear (割創), open spear (裂創), stent (射創), and gallows (咬創) according to the shape of the wound.創) is divided into plate-shaped window (瓣狀創) defect window. In addition, in relation to the body cavity (体腔), it is divided into a common sore (穿通創) and a non-heavy sore, and an infected and non-infectious sore according to the presence or absence of infection. Symptoms associated with the wound include pain, bleeding, and dysfunction. When you get hurt, your body reacts quickly and tries to recover quickly from the injury (refer to Doosan Encyclopedia).

The healing process varies depending on the general condition of the living body or the wound, but it is usually healed according to the following process. ① Degeneration and death of injured cells, ② Migrating cells from surrounding tissues and leakage of tissue fluid, ③ Precipitation of fibrin and formation of granulation tissue.

In order to heal a wound quickly, a treatment method according to the wound must be applied. However, there are cases in which the wound healing is delayed or scars remain because the patient lacks knowledge about wound treatment. For example, tincture of iodine or Mercurochrome, called red pills, are commonly used for wound healing. Tincture of iodine or Mercurochrome is an antibacterial agent that kills bacteria. However, these drugs have a sterilizing effect, but in the process of actually healing any wound, they interfere with the epithelial cells or fibroblasts in the regeneration process, thereby slowing the healing process.

Also, in the case of burns, the evaluation method of the depth of damage caused by burns has not been studied much compared to its importance. In addition, the judgment is highly dependent on the subjective judgment by experts. In addition, there is a shortage of image experts worldwide. If the general public underestimates the initial damage caused by burns without consulting a burn specialist, the wound may get worse, and if the initial damage depth is overestimated, it may cause cost increases.

Recently, deep learning has been showing good performance in the image field. In particular, image classification through CNN algorithm is being actively studied in the medical field. If the initial damage can be determined relatively accurately by creating a model that can identify the depth of burn damage through the CNN algorithm using the digital image, it prevents the patient from spending unnecessary money or the patient's wound from getting worse. can do. In particular, early judgment of the depth of burn damage in medically vulnerable areas such as islands and mountains, schools, and military bases can prevent aggravation of wounds and prevent unnecessary medical expenses.

The operating method of the wound management apparatus of the present disclosure includes: acquiring image data of a plurality of existing wounds; acquiring treatment information of existing wounds corresponding to image data of a plurality of existing wounds; and image data of a plurality of existing wounds and acquiring a machine learning model by learning the correlation between the treatment information of the existing wound with CNN (Convolution Neural Network), obtaining the wound image data of the current patient, and applying the machine learning model to the wound image data of the current patient determining the wound treatment expected information, and outputting the wound treatment expected information, wherein the existing wound treatment information includes wound depth information, wound treatment period information, surgery information, and scarring information. All of the information is included, and the wound treatment prediction information is characterized in that it includes expected information on the presence or absence of surgery and the expected information on the presence or absence of scarring.

In the method of operation of the wound management apparatus of the present disclosure, the wound depth information selects one of "first-degree burns", "second-degree burns superficially", "second-degree burns deep", "third-degree burns", and "fourth-degree burns" and the wound healing period information represents the number of days from the date of injury to the completion date of re-epithelialization, and the step of determining the wound treatment expected information is the depth of the wound obtained by applying a machine learning model to the wound image data of the current patient. When the information indicates "first-degree burn" or "second-degree burn superficiality", information (a) on whether or not the first scar has occurred is set to 0, and the obtained wound depth information is "second-degree burn deep", " Step of setting the information (a) on the presence or absence of the first scar to 1 when indicating a third-degree burn or a "fourth-degree burn" When the information is "7 days" or less, the information (b) on the occurrence of the second scar is set to 0, and when the information about the treatment period of the obtained wound exceeds "7 days", information on the presence or absence of the second scar ( b) is set to 1, when the information on whether scarring obtained by applying a machine learning model to the current patient's wound image data indicates that there is no scar left, the third information on whether scarring occurs (c) is set to 0 setting, and when the obtained information on whether scars remain scars, setting the third information (c) on whether or not scars occur to 1, based on the predetermined weights (w1, w2, w3), the first If the weighted average ((w1*a+w2*b+w3*c)/3) of the information on the presence or absence of scarring or the information on the presence or absence of the third scar is greater than 0.5, the expected information on the presence or absence of scarring When the weighted average ((w1*a+w2*b+w3*c)/3) of the information on the presence or absence of the first scar and the information on the presence or absence of the third scar is 0.5 or less, scarring Including the step of determining the expected information on the presence or absence of a scar as not leaving a scar, w3 is greater than w1 and w2 It is large and characterized in that w1+w2+w3<=3.

In the method of operating the wound management apparatus of the present disclosure, the determining of the wound treatment prediction information includes determining that the wound depth information obtained by applying a machine learning model to the wound image data of the current patient is “first-degree image” or “second-degree image”. In the case of showing “superficial burns”, the information (d) on the presence or absence of the first operation is set to 0, and when the acquired wound depth information shows “second-degree burns,” the information (d) on the presence or absence of the first surgery is set to 1 /2, and when the acquired wound depth information indicates "third-degree burn" or "fourth-degree burn", setting the information (d) on the presence or absence of the first operation to 1, in the current patient's wound image data If the information on whether or not surgery obtained by applying the machine learning model indicates "no need for surgery", the information on whether or not the second operation (e) is set to 0, and the information on whether or not the operation is obtained is "required for surgery" ", setting the information (e) on the presence or absence of the second operation to 1, based on the predetermined weights (w4, w5), the information (d) on the presence or absence of the first operation and the presence or absence of the second operation When the weighted average ((w4*d+w5*e)/2) of the information (e) for the information (e) is greater than 0.5, the step of determining the expected information on the presence or absence of surgery as “needs surgery”, and the If the weighted average ((w4*d+w5*e)/2) of information (d) and information (e) on the presence or absence of the second surgery is less than 0.5, the expected information on the presence or absence of surgery is determined as “no surgery required” It is characterized in that w5 is greater than w4, and w4+w5<=2.

The operation method of the wound management apparatus of the present disclosure includes the steps of obtaining hospital information, outputting the obtained hospital information, and receiving reservation information related to a hospital reservation from a user when the expected information on the presence or absence of scarring is set to require surgery. Receiving, transmitting information about the patient and reservation information to the hospital server, and further comprising the steps of receiving a reservation confirmation signal from the hospital server, the reservation information is the patient's scheduled visit time and the patient's preferred healthcare provider including at least one of information about the hospital, and the hospital information includes contact information of the hospital, the location of the hospital, the distance between the hospital and the patient, transportation to the hospital, business hours of the hospital, information of medical personnel belonging to the hospital, the opening date of the hospital, the hospital of patients, the number of times of reuse, the rating of the hospital, the current number of waiting at the hospital, parking space near the hospital, and a description of the hospital facilities further include at least one of, and the treatment information of the existing wound includes the patient's age information, the patient's gender information, It further includes wound healing period information, information on the number of visits to wound healing, information on the degree of scarring and cost information until healing. It is characterized in that it further includes information about the degree and the estimated cost of healing.

In the wound management apparatus of the present disclosure, the wound management apparatus includes a processor and a memory, and the processor, based on an instruction stored in the memory, obtains image data of a plurality of existing wounds, corresponding to the image data of the plurality of existing wounds Acquiring treatment information of an existing wound that becomes The steps of acquiring data, applying a machine learning model to the current patient's wound image data to determine wound treatment prediction information, and outputting wound treatment prediction information are performed, and the treatment information of the existing wound is the depth of the wound. Information, wound treatment period information, information on whether or not surgery, and information on whether or not scarring is included, and the wound treatment expected information includes expected information on whether or not surgery and predictive information on whether scars occur. .

The wound depth information of the wound management device of the present disclosure indicates one of “first-degree burns”, “second-degree burns superficial”, “second-degree burns deep”, “third-degree burns” and “fourth-degree burns”, and treatment of the wound The period information represents the number of days from the date of injury to the completion of re-epithelialization, and the processor determines the wound treatment expected information, and the depth information of the wound obtained by applying a machine learning model to the wound image data of the current patient is "1". In the case of showing "degree burn" or "second-degree burn superficiality", the information (a) on whether or not the first scar has occurred is set to 0, and the obtained wound depth information is "second-degree burn deep", "third-degree burn" Alternatively, if “4th degree burn” is indicated, the step of setting the information (a) on the presence or absence of the first scar to 1, the treatment period information of the wound obtained by applying the machine learning model to the wound image data of the current patient is “7” If less than one day, the information (b) on the occurrence of the second scar is set to 0, and when the acquired wound treatment period information is more than “7 days”, the information (b) on the presence or absence of the second scar is set to 1 When the information on whether scars obtained by applying a machine learning model to the current patient's wound image data indicates that no scars remain, the third information (c) on whether or not scars occur is set to 0, and acquired When the information on the presence of scarring indicates that scars remain, the step of setting the information (c) on whether or not the third scar is generated to 1, based on the predetermined weights (w1, w2, w3), the presence or absence of the first scar When the weighted average ((w1*a+w2*b+w3*c)/3) of the information on scarring or the third scarring is greater than 0.5, it is the When the weighted average ((w1*a+w2*b+w3*c)/3) of information on the stage, and information on whether or not the first scar has occurred to the information on whether or not the third scar is generated is 0.5 or less, performing the step of determining the expected information as non-scarring, w3 is greater than w1 and w2, It is characterized in that w1+w2+w3<=3.

The processor of the wound management device of the present disclosure sets the wound depth information obtained by applying a machine learning model to the current patient's wound image data to determine the wound treatment prediction information as "first-degree burn" or "second-degree image superficiality" In the case of showing , the information (d) on the presence or absence of the first operation is set to 0, and when the acquired wound depth information indicates "second-degree burn deepness", the information (d) on the presence or absence of the first operation is halved. setting, and setting the information (d) on the presence or absence of the first operation to 1 when the acquired wound depth information indicates "third-degree burn" or "fourth-degree burn"; If the information on whether surgery obtained by applying In this case, the step of setting the information (e) on the presence or absence of the second operation to 1, the information on the presence or absence of the first operation (d) and the information on the presence or absence of the second operation ( If the weighted average of e) ((w4*d+w5*e)/2) is greater than 0.5, determining the expected information on the presence or absence of surgery as “needs surgery”, and information on the presence or absence of the first operation (d) ) and the weighted average ((w4*d+w5*e)/2) of information (e) on the presence or absence of the second operation is 0.5 or less, the step of determining the expected information on the presence or absence of surgery as “no surgery required” It is characterized in that w5 is greater than w4, and w4+w5<=2.

The processor of the wound management device of the present disclosure obtains hospital information, outputting the obtained hospital information, when the expected information on the presence or absence of scarring is set as a need for surgery, based on the instruction stored in the memory, from the user to the hospital Receiving the reservation information related to the reservation, transmitting information and reservation information about the patient to the hospital server, and further performing the steps of receiving a reservation confirmation signal from the hospital server, the reservation information is the patient's scheduled visit time and The patient includes at least one of information about a preferred medical provider, and the hospital information includes contact information of the hospital, the location of the hospital, the distance between the hospital and the patient, transportation to the hospital, business hours of the hospital, information of medical personnel belonging to the hospital, It further includes at least one of the opening date of the hospital, the number of re-uses of the hospital's patients, the rating of the hospital, the current number of waiting people in the hospital, the parking space near the hospital, and a description of the hospital facilities, and the treatment information of the existing wound is the patient's age information , the patient's gender information, wound healing period information, number of visits to wound healing, information on the degree of scarring, and cost information until healing. It is characterized in that it further includes information on the number of times, information on the expected degree of scarring, and information on the estimated cost of healing.

In addition, a program for implementing the method of operation of the wound management apparatus as described above may be recorded in a computer-readable recording medium.

1 is a block diagram of a wound management apparatus 100 according to an embodiment of the present disclosure.
2 is a view showing a wound management system according to an embodiment of the present disclosure.
3 is a flowchart illustrating the operation of the wound management apparatus according to an embodiment of the present disclosure.
4 is a block diagram illustrating a wound management apparatus according to an embodiment of the present disclosure.
5 is a block diagram illustrating a wound management apparatus according to an embodiment of the present disclosure.
6 is a flowchart illustrating a process of acquiring expected information on the presence or absence of surgery according to an embodiment of the present disclosure.
7 is a flowchart illustrating a process of acquiring expected information on whether scars occur according to an embodiment of the present disclosure.
8 is a diagram illustrating a user terminal according to an embodiment of the present disclosure.

Advantages and features of the disclosed embodiments, and methods of achieving them, will become apparent with reference to the embodiments described below in conjunction with the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only the present embodiments allow the present disclosure to be complete, and those of ordinary skill in the art to which the present disclosure pertains. It is only provided to fully inform the person of the scope of the invention.

Terms used in this specification will be briefly described, and the disclosed embodiments will be described in detail.

Terms used in this specification have been selected as currently widely used general terms as possible while considering the functions in the present disclosure, but these may vary depending on the intention or precedent of a person skilled in the art, the emergence of new technology, and the like. In addition, in a specific case, there is a term arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the description of the corresponding invention. Therefore, the terms used in the present disclosure should be defined based on the meaning of the term and the contents of the present disclosure, rather than the simple name of the term.

References in the singular herein include plural expressions unless the context clearly dictates the singular. Also, the plural expression includes the singular expression unless the context clearly dictates the plural.

In the entire specification, when a part "includes" a certain element, this means that other elements may be further included, rather than excluding other elements, unless otherwise stated.

Also, as used herein, the term “unit” refers to a software or hardware component, and “unit” performs certain roles. However, "part" is not meant to be limited to software or hardware. A “unit” may be configured to reside on an addressable storage medium and may be configured to refresh one or more processors. Thus, by way of example, “part” includes components such as software components, object-oriented software components, class components and task components, processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. The functionality provided within components and “parts” may be combined into a smaller number of components and “parts” or further divided into additional components and “parts”.

According to an embodiment of the present disclosure, “unit” may be implemented with a processor and a memory. The term “processor” should be interpreted broadly to include general purpose processors, central processing units (CPUs), microprocessors, digital signal processors (DSPs), controllers, microcontrollers, state machines, and the like. In some circumstances, a “processor” may refer to an application specific semiconductor (ASIC), a programmable logic device (PLD), a field programmable gate array (FPGA), or the like. The term "processor" refers to a combination of processing devices, such as, for example, a combination of a DSP and a microprocessor, a combination of a plurality of microprocessors, a combination of one or more microprocessors in combination with a DSP core, or any other such configurations. may refer to.

The term “memory” should be interpreted broadly to include any electronic component capable of storing electronic information. The term memory includes random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), programmable read-only memory (PROM), erase-programmable read-only memory (EPROM), electrical may refer to various types of processor-readable media, such as erasable PROM (EEPROM), flash memory, magnetic or optical data storage, registers, and the like. A memory is said to be in electronic communication with the processor if the processor is capable of reading information from and/or writing information to the memory. A memory integrated in the processor is in electronic communication with the processor.

Hereinafter, with reference to the accompanying drawings, embodiments will be described in detail so that those of ordinary skill in the art to which the present disclosure pertains can easily implement them. And in order to clearly describe the present disclosure in the drawings, parts not related to the description will be omitted.

1 is a block diagram of a wound management apparatus 100 according to an embodiment of the present disclosure.

Referring to FIG. 1 , the wound management apparatus 100 according to an embodiment may include a data learning unit 110 and a data recognition unit 120 . The wound management apparatus 100 as described above may include a processor and a memory.

The data learning unit 110 may receive the first data and the second data to obtain a machine learning model for the relationship between the first data and the second data. The machine learning model obtained by the data learning unit 110 may be a model for generating second data using the first data. For example, the data learning unit 110 may learn a relationship between image data of a plurality of existing wounds and treatment information on the existing wounds. When an image of a wound and basic information about the wound are given, the data learning unit 110 may generate a machine learning model for predicting treatment information according to the wound.

The data recognition unit 120 may output wound treatment prediction information by applying image data of the current patient's wound or information related to the wound to the machine learning model. The machine learning model may be information about the criteria for which treatment should be performed according to the image of the wound. In addition, the data recognition unit 120 may use the image data of the wound or information related to the wound and the result output by the machine learning model to update the machine learning model.

At least one of the data learning unit 110 and the data recognition unit 120 may be manufactured in the form of at least one hardware chip and mounted in an electronic device. For example, at least one of the data learning unit 110 and the data recognition unit 120 may be manufactured in the form of a dedicated hardware chip for artificial intelligence (AI), or a conventional general-purpose processor (eg, CPU Alternatively, it may be manufactured as a part of an application processor) or a graphics-only processor (eg, GPU) and mounted on various electronic devices described above.

Also, the data learning unit 110 and the data recognition unit 120 may be respectively mounted on separate electronic devices. For example, one of the data learning unit 110 and the data recognition unit 120 may be included in the electronic device, and the other may be included in the server. In addition, the data learning unit 110 and the data recognition unit 120 may provide the machine learning model information built by the data learning unit 110 to the data recognition unit 120 through wire or wireless, and recognize data. Data input to the unit 120 may be provided to the data learning unit 110 as additional learning data.

Meanwhile, at least one of the data learning unit 110 and the data recognition unit 120 may be implemented as a software module. When at least one of the data learning unit 110 and the data recognition unit 120 is implemented as a software module (or a program module including instructions), the software module is a memory or computer-readable ratio It may be stored in a non-transitory computer readable media. Also, in this case, at least one software module may be provided by an operating system (OS) or may be provided by a predetermined application. Alternatively, a part of the at least one software module may be provided by an operating system (OS), and the other part may be provided by a predetermined application.

The data learning unit 110 according to an embodiment of the present disclosure includes a data acquisition unit 111 , a preprocessing unit 112 , a training data selection unit 113 , a model learning unit 114 , and a model evaluation unit 115 . may include

The data acquisition unit 111 may acquire data necessary for machine learning. Since a lot of data is required for learning, the data acquisition unit 111 may receive image data of a plurality of existing wounds or information on wounds and treatment information of existing wounds.

The preprocessor 112 may preprocess the obtained data so that the received data can be used for machine learning. The preprocessor 112 may process the obtained data into a preset format so that the model learning unit 114 to be described later can use it. The preprocessor 112 may process the wound image data with a filter to make the outline of the wound clearly visible. In addition, the preprocessor 112 may adjust at least one of the size, white balance, color, contrast, or saturation of the image data so that the characteristics of the wound are clearly revealed.

The learning data selection unit 113 may select data necessary for learning from among the preprocessed data. The selected data may be provided to the model learning unit 114 . The learning data selection unit 113 may select data necessary for learning from preprocessed data according to a preset criterion. In addition, the training data selection unit 113 may select data according to a preset criterion by learning by the model learning unit 114 to be described later.

The training data selection unit 113 may pre-classify the image data based on wound information or a pre-trained machine learning model in order to relieve the processing burden of the model learning unit 114 . The learning data selection unit 113 may classify image data of an existing wound according to at least one of a type and a size of the wound based on the information on the wound. Different treatment methods will be used depending on whether the wound is a burn or a cut, and the preprocessor 112 may classify only the image of the burn wound separately. Since the model learning unit 114 will perform machine learning specialized for image wounds, the processing burden may be reduced.

The model learning unit 114 may learn a criterion regarding which wound treatment prediction information to output according to the image data of the existing wound. In addition, the model learning unit 114 may learn by using a data learning model for outputting treatment prediction information of a wound according to image data of the existing wound and information about the wound as learning data. In this case, the data learning model may be a pre-built model. For example, the data learning model may be a model built in advance by receiving basic learning data (eg, a wound image, information about a wound, or treatment information of a wound).

The data learning model may be constructed in consideration of the field of application of the learning model, the purpose of learning, or the computer performance of the device. The data learning model may be, for example, a model based on a neural network. For example, models such as Deep Neural Network (DNN), Recurrent Neural Network (RNN), Long Short-Term Memory models (LSTM), BRDNN (Bidirectional Recurrent Deep Neural Network), Convolutional Neural Networks (CNN) can be used as data learning models. However, it is not limited thereto.

According to various embodiments, the model learning unit 114 may determine, as a data learning model to learn, a data learning model having a high correlation between the input learning data and the basic learning data when there are a plurality of pre-built data learning models. have. In this case, the basic learning data may be pre-classified for each type of data, and the data learning model may be previously built for each type of data. For example, the basic training data may be pre-classified according to various criteria such as a place where the training data is generated, a time when the training data is generated, the size of the training data, a generator of the training data, and the type of object in the training data.

Also, the model learning unit 114 may train the data learning model using, for example, a learning algorithm including error back-propagation or gradient descent.

Also, the model learning unit 114 may learn the data learning model through supervised learning using, for example, learning data as an input value. In addition, the model learning unit 114 learns data, for example, through unsupervised learning that discovers a criterion for situation determination by self-learning the type of data required for situation determination without any guidance. model can be trained. Also, the model learning unit 114 may learn the data learning model through, for example, reinforcement learning using feedback on whether the result of the situation determination according to the learning is correct.

Also, when the data learning model is learned, the model learning unit 114 may store the learned data learning model. In this case, the model learning unit 114 may store the learned data learning model in the memory of the electronic device including the data recognition unit 120 . Alternatively, the model learning unit 114 may store the learned data learning model in a memory of a server connected to the electronic device through a wired or wireless network.

The memory in which the learned data learning model is stored may also store, for example, commands or data related to at least one other component of the electronic device. The memory may also store software and/or programs. A program may include, for example, a kernel, middleware, an application programming interface (API) and/or an application program (or "application"), and the like.

The model evaluation unit 115 inputs evaluation data to the data learning model, and when a result output from the evaluation data does not satisfy a predetermined criterion, the model evaluation unit 114 may allow the model learning unit 114 to learn again. In this case, the evaluation data may be preset data for evaluating the data learning model.

For example, the model evaluation unit 115, among the results of the learned data learning model for the evaluation data, does not satisfy a predetermined criterion when the number or ratio of evaluation data for which the recognition result is not accurate exceeds a preset threshold. can be evaluated as For example, when the predetermined criterion is defined as a ratio of 2%, when the learned data learning model outputs an erroneous recognition result for more than 20 evaluation data out of a total of 1000 evaluation data, the model evaluation unit 115 learns It can be evaluated that the existing data learning model is not suitable.

On the other hand, when there are a plurality of learned data learning models, the model evaluation unit 115 evaluates whether each of the learned video learning models satisfies a predetermined criterion, and uses the model that satisfies the predetermined criterion as the final data learning model. can decide In this case, when there are a plurality of models satisfying the predetermined criteria, the model evaluator 115 may determine any one or a predetermined number of models preset in the order of the highest evaluation score as the final data learning model.

Meanwhile, at least one of the data acquisition unit 111 , the preprocessor 112 , the training data selection unit 113 , the model learning unit 114 , and the model evaluation unit 115 in the data learning unit 110 is at least one It may be manufactured in the form of a hardware chip of For example, at least one of the data acquisition unit 111 , the preprocessor 112 , the training data selection unit 113 , the model learning unit 114 , and the model evaluation unit 115 is artificial intelligence (AI). It may be manufactured in the form of a dedicated hardware chip for this purpose, or it may be manufactured as a part of an existing general-purpose processor (eg, CPU or application processor) or graphics-only processor (eg, GPU) and mounted on the various electronic devices described above.

In addition, the data acquisition unit 111 , the preprocessor 112 , the training data selection unit 113 , the model learning unit 114 , and the model evaluation unit 115 may be mounted in one electronic device, or separate Each of the electronic devices may be mounted. For example, some of the data acquisition unit 111 , the preprocessor 112 , the training data selection unit 113 , the model learning unit 114 , and the model evaluation unit 115 are included in the electronic device, and the rest are It can be included in the server.

In addition, at least one of the data acquisition unit 111 , the preprocessor 112 , the training data selection unit 113 , the model learning unit 114 , and the model evaluation unit 115 may be implemented as a software module. A program in which at least one of the data acquisition unit 111, the preprocessor 112, the learning data selection unit 113, the model learning unit 114, and the model evaluation unit 115 includes a software module (or instructions) module), the software module may be stored in a computer-readable non-transitory computer readable medium. Also, in this case, at least one software module may be provided by an operating system (OS) or may be provided by a predetermined application. Alternatively, a part of the at least one software module may be provided by an operating system (OS), and the other part may be provided by a predetermined application.

The data recognition unit 120 according to an embodiment of the present disclosure includes a data obtaining unit 121 , a preprocessing unit 122 , a recognition data selecting unit 123 , a recognition result providing unit 124 , and a model updating unit 125 . may include

The data acquisition unit 121 may acquire wound image data of the current patient or information related to the wound of the current patient. The preprocessor 122 may preprocess the acquired data so that the acquired current patient's wound image data and information related to the current patient's wounds can be used. For example, the pre-processing unit 122 may process the wound image data with a filter, so that the outline of the wound is clearly revealed. In addition, the preprocessor 122 may adjust at least one of white balance, color, contrast, or saturation of the wound image data.

The recognition data selection unit 123 may select necessary data from the pre-processed data. The selected data may be provided to the recognition result providing unit 124 . The recognition data selection unit 123 may select some or all of the preprocessed data according to a preset criterion. Also, the recognition data selection unit 123 may select data according to a preset criterion by learning by the model learning unit 114 .

The recognition result providing unit 124 may obtain result data by applying the selected data to the data learning model. The recognition result providing unit 124 may output the result data as voice or text. The recognition result providing unit 124 may apply the selected data to the data learning model by using the data selected by the recognition data selecting unit 123 as an input value. Also, the recognition result may be determined by a data learning model.

The model updating unit 125 may update the data learning model based on the evaluation of the recognition result provided by the recognition result providing unit 124 . For example, the model updating unit 125 may provide the model learning unit 114 with the recognition result provided by the recognition result providing unit 124 so that the model learning unit 114 updates the data learning model. have.

Meanwhile, at least one of the data acquisition unit 121 , the preprocessor 122 , the recognition data selection unit 123 , the recognition result providing unit 124 , and the model update unit 125 in the data recognition unit 120 is at least It may be manufactured in the form of a single hardware chip and mounted in an electronic device. For example, at least one of the data acquisition unit 121 , the preprocessor 122 , the recognition data selection unit 123 , the recognition result providing unit 124 , and the model update unit 125 is artificial intelligence (AI). ), or may be manufactured as a part of an existing general-purpose processor (eg, CPU or application processor) or graphics-only processor (eg, GPU) and mounted on the various electronic devices described above.

In addition, the data acquisition unit 121 , the preprocessor 122 , the recognition data selection unit 123 , the recognition result providing unit 124 , and the model update unit 125 may be mounted in one electronic device or separately. It may be mounted on each of the electronic devices of For example, some of the data acquiring unit 121 , the preprocessing unit 122 , the recognition data selection unit 123 , the recognition result providing unit 124 , and the model updating unit 125 are included in the electronic device, and the remaining parts are included in the electronic device. may be included in the server.

In addition, at least one of the data acquisition unit 121 , the preprocessor 122 , the recognition data selection unit 123 , the recognition result providing unit 124 , and the model update unit 125 may be implemented as a software module. At least one of the data acquisition unit 121, the preprocessor 122, the recognition data selection unit 123, the recognition result providing unit 124, and the model update unit 125 includes a software module (or instructions) When implemented as a program module), the software module may be stored in a computer-readable non-transitory computer readable medium. Also, in this case, at least one software module may be provided by an operating system (OS) or may be provided by a predetermined application. Alternatively, a part of the at least one software module may be provided by an operating system (OS), and the other part may be provided by a predetermined application.

2 is a view showing a wound management system according to an embodiment of the present disclosure.

The wound management system 200 may include an input unit 210 , a wound management device 220 , and an output unit 230 .

The input unit 210 may be included in the user terminal. The input unit 210 may acquire image data of the wound. In addition, the input unit 210 may obtain information related to the image of the wound. The image data of the wound may be an image of the wound directly taken by the user using a camera. In addition, the image data of the wound may be image data received by the user by wire or wirelessly from another user terminal. In addition, the image data of the wound may be image data already stored in the user terminal.

The wound-related information may include at least one of a wounded body part corresponding to the image data of the wound, the cause of the wound, the date of occurrence of the wound, the size of the wound, the patient's age, the patient's skin color, and the patient's gender. have.

The input unit 210 may transmit image data of the wound or information related to the wound to the wound management apparatus 220 through wired or wireless. The input unit 210 may compress image data of a wound or information related to a wound based on a predetermined algorithm in order to reduce the amount of data.

Wound management device 220 may be included in the user terminal or server. Wound management device 220 may perform machine learning based on the data received from the input unit 210, or may output the result data based on the machine learning model already machine learning. The wound management device 220 may output a machine learning model based on the received image data of the wound or information related to the wound. Also, the wound management apparatus 220 may output wound treatment prediction information by applying the received image data of the wound or information related to the wound to the machine learning model.

The wound management device 220 may transmit the generated machine learning model or wound treatment prediction information to the output unit 230 through a wired or wireless connection. The wound management device 220 may compress the generated machine learning model or wound treatment prediction information using a predetermined algorithm, thereby reducing the amount of data required for data transmission and increasing the transmission speed.

The output unit 230 may correspond to a user terminal. The output unit 230 may be included in the same user terminal as the input unit 210 . The output unit 230 may apply wound image data or wound-related information stored in the memory to the machine learning model to output wound treatment prediction information. Also, the output unit 230 may output the wound treatment prediction information received from the wound management apparatus 220 . The output unit 230 may output the result to the display unit or may output the result in sound through a speaker. The user can view or hear the wound treatment prediction information and take necessary actions. Therefore, the user can prevent the case of aggravation of the wound or the case that the cost is excessively spent.

Hereinafter, the operation of the wound management device 220 will be described in more detail.

3 is a flowchart illustrating the operation of the wound management apparatus according to an embodiment of the present disclosure.

The wound management apparatus 220 may perform step 310 of acquiring image data of a plurality of existing wounds. The image data of the plurality of existing wounds may be received from the input unit 210 .

The image data of a plurality of existing wounds may further include an image of a color patch as well as an image of the wound. The image of the color patch is an image in which at least one color is displayed in different regions. When the user places a color patch next to the wound and takes an image of the wound, the image data of the existing wound may include the color patch image. The preprocessor 112 of the wound management apparatus 220 may correct the image data of the existing wound based on the color patch image included in the image data of the existing wound. The color patch image included in the image data of the existing wound is a photograph of the same color patch. Therefore, the wound management device 220 may correct the image data of the existing wound based on the color patch, thereby minimizing the image caused by the lighting when photographing the wound. Also, the wound management apparatus 220 may correct at least one of white balance, brightness, and saturation of image data of an existing wound based on the color patch. Since the effect of lighting on the existing wound image data is minimized through correction, the wound management device 220 can efficiently perform machine learning.

The image data of the plurality of existing wounds may be image data obtained from the wounds of the existing patient. The image data of the plurality of existing wounds may be images of wounds taken by medical personnel while treating patients. A medical practitioner may input various information related to the image data to the wound management apparatus 220 together with image data of a plurality of existing wounds. The image data of the plurality of existing wounds may be image data arranged in chronological order from a wound occurrence date until the wound is completely healed. In addition, the image data of a plurality of existing wounds may be image data in which the wound site is clearly revealed so that it can be used for machine learning. The user may select image data that can be used for machine learning from among the plurality of image data of existing wounds.

The wound management apparatus 220 may perform an operation 320 of obtaining treatment information of the existing wounds corresponding to the image data of the plurality of existing wounds. The treatment information of the existing wound may include a diagnosis result of at least one wound care professional.

The wound management apparatus 220 may obtain wound treatment information based on the patient's medical record corresponding to the wound image data. In addition, the wound management apparatus 220 may acquire treatment information of the wound based on the result of the existing diagnostic equipment such as laser Doppler.

Existing wound treatment information may include all of the information on the depth of the wound, the healing period of the wound, and whether the wound is operated on or a scar is generated. Existing wound treatment information further includes at least one of the patient's age, gender, number of visits to healing, whether surgery, time required for surgery, post-operative healing period, information on the degree of scarring, or information on the cost of healing can do.

According to an embodiment of the present disclosure, the information on the depth of the wound is information indicating how severe the wound is based on a predetermined criterion by a medical practitioner. If the wound is a burn, the healthcare provider can determine the severity of the wound as one of a first-degree burn, a second-degree burn, or a third-degree burn. The wound management apparatus 220 may store information on the depth of the wound corresponding to the image data as one of a “first-degree image”, a “second-degree image” or a “third-degree image”.

According to another embodiment of the present disclosure, the information on the depth of the wound is information indicating how severe the wound is based on a predetermined criterion by a medical practitioner. If the wound is a burn, the practitioner can determine the severity of the wound as one of a first-degree burn, a second-degree burn superficial, a second-degree burn deep, a third-degree burn, or a fourth-degree burn. The wound management device 220 stores information on the depth of the wound corresponding to the image data as “first-degree burn”, “second-degree burn superficial”, “second-degree burn deep”, “third-degree burn” or “fourth-degree burn”. You can save one of them.

The wound management apparatus 220 may store the existing wound treatment information corresponding to the image data of the existing wound as an index. For example, the wound management apparatus 220 may indicate an image as “0” and a cut as “1”. In addition, the wound management apparatus 220 may represent the first degree image as "00", the second degree image as "01", and the third degree image as "02".

Also, when the wound is a burn, the depth of the wound can be further subdivided. For example, a practitioner may determine the severity of the wound as one of a first-degree burn, a second-degree burn superficial, a second-degree burn deep, a third-degree burn, or a fourth-degree burn. The wound management device 220 stores information on the depth of the wound corresponding to the image data as “first-degree burn”, “second-degree burn superficial”, “second-degree burn deep”, “third-degree burn” or “fourth-degree burn”. You can save one of them. Here, "first degree burn" means damage to the epidermis or keratin of the skin, "superficial second degree burn" means damage to the epidermis or shallow dermal layer of the skin, and "second degree burns" means damage to the deep dermal layer of the skin means that In addition, "third degree burn" means damage to the entire dermal layer and subcutaneous tissue, and "fourth degree burn" means damage to the entire dermal layer, subcutaneous tissue, fat, muscle, and bone.

The wound management apparatus 220 may store the existing wound treatment information corresponding to the image data of the existing wound as an index. For example, the wound management apparatus 220 may indicate an image as “0” and a cut as “1”. In addition, the wound care device 220 denotes the first-degree image by "00", the second-degree image superficiality by "01", the second-degree image deepness by "02", the third-degree image by "03", A fourth degree image can be represented by "04".

The wound healing period may be information indicating the number of days from the occurrence of the wound to the completion of re-epithelialization. The wound healing period may be divided into predetermined groups. For example, if 0-7 days are required to heal the wound shown in the existing wound image data, the wound management device 220 classifies the existing wound image data into a first group, and if 8-14 days are required to heal the wound, the second group It can be divided into groups, and if it takes 15-21 days to heal a wound, it can be divided into a 3rd group, and when it takes more than 22 days, it can be divided into a 4th group.

Whether the wound has been operated on may be information indicating whether a medical professional has operated on the wound of an existing patient. For example, the wound treatment apparatus 220 may store "1" as information on whether or not the operation was performed if the medical personnel performed an operation for wound treatment, and store "0" as information on whether or not the operation was performed if no operation was performed. .

In addition, the information on whether scars have occurred may be information on whether or not a scar has occurred after the medical personnel completes the treatment. For example, if a scar has occurred, the wound management device 220 stores “1” as information on whether or not a scar has occurred, and if no scar has occurred, the wound management device 220 stores “0” as information about whether a scar has occurred can be saved.

The wound management apparatus 220 may perform a step 330 of acquiring a machine learning model by learning the correlation between image data of a plurality of existing wounds and treatment information of the existing wounds using a Convolution Neural Network (CNN). This will be described in more detail with reference to FIG. 4 .

CNN is a type of deep neural network (DNN), and is a neural network composed of one or several convolutional layers, a pooling layer, and fully connected layers. CNN has a structure suitable for learning two-dimensional data, and can be trained through a backpropagation algorithm. It is one of the representative models of DNN that is widely used in various application fields such as object classification in images and object detection.

4 is a block diagram illustrating a wound management apparatus according to an embodiment of the present disclosure.

The wound management device 220 may include a data learning unit 410 . The data learner 410 may correspond to the data learner 110 of FIG. 1 .

The data learning unit 410 may receive image data 421 of a plurality of existing wounds. In addition, the data learning unit 410 may receive the treatment information 422 of the existing wound corresponding to each of the image data of the plurality of existing wounds. Treatment information for existing wounds can be prepared by a medical professional.

The data learning unit 410 may learn the relationship between the image data 421 of the wound and the treatment information 422 of the wound and output it to the machine learning model 431 . The wound management apparatus 220 may separately acquire the machine learning model 431 for each of the wound depth information, the wound treatment period information, the operation information, and the scarring information.

The data learning unit 410 may learn the relationship between "wound-related information and wound image data 421 ," and wound treatment information 422 , and output it to the machine learning model 431 . The wound-related information may include at least one of a cause of a wound corresponding to the image data of the wound, a date of occurrence of the wound, a size of a wound, an age of a patient, a skin color of a patient, and a gender of the patient.

The wound management device 220 may store the machine learning model 431 . Also, the wound management apparatus 220 may transmit the machine learning model 431 to another wound management apparatus. The wound management apparatus 220 may obtain wound treatment prediction information by applying the image data of the new wound to the machine learning model 431 . Wound treatment prediction information is described in detail below.

Referring back to FIG. 3 , the wound management apparatus 220 may perform step 340 of acquiring wound image data of the current patient. The user may photograph the wound of the patient through the camera included in the user terminal and transmit it to the wound management apparatus 220 . The user may transmit the image data already stored in the memory to the wound management device 220 . The user may transmit image data received from another terminal to the wound management apparatus 220 .

The current patient's wound image data may further include an image of a color patch as well as an image of the wound. When the user places a color patch next to the wound of the current patient and takes an image of the wound, the wound image data of the current patient may include the color patch image. The preprocessor 122 of the wound management device 220 may correct the wound image data of the current patient based on the color patch image included in the wound image data of the current patient. The wound management device 220 may correct the wound image data of the current patient based on the color patch, thereby minimizing the image caused by the lighting when photographing the wound. In addition, the wound management apparatus 220 may correct at least one of white balance, brightness, and saturation of the current patient's wound image data based on the color patch. Since the effect of lighting on the current patient's wound image data is minimized through correction, the wound management device 220 can obtain the best prediction result by applying the machine learning model to the current patient's wound image data.

As already described, the wound management device 220 may be included in a server or a user terminal. The wound management apparatus 220 may perform the step 350 of determining wound treatment prediction information by applying a machine learning model to the wound image data of the current patient. In addition, the wound management apparatus 220 may perform an operation 360 of outputting the wound treatment prediction information. Hereinafter, a process in which the wound management device 220 acquires wound treatment prediction information using a machine learning model will be described in more detail with reference to FIG. 5 .

5 is a block diagram illustrating a wound management apparatus according to an embodiment of the present disclosure.

The wound management device 220 may include the data recognition unit of FIG. 1 . The data recognition unit may obtain result data by applying the input data to the machine learning model 510 obtained in advance. Specifically, the wound management device 220 may obtain the wound treatment prediction information 530 by applying the wound image data 520 of the current patient to the machine learning model 510 .

The wound management device 220 may obtain the wound treatment prediction information 530 by applying the “wound-related information” and the wound image data 520 of the current patient to the machine learning model 510 . The wound-related information may include at least one of a cause of a wound corresponding to the image data of the wound, a date of occurrence of the wound, a size of a wound, an age of a patient, a skin color of a patient, and a gender of the patient.

The wound treatment prediction information may include information about the depth of the wound, information about the treatment period of the wound, information about whether surgery is performed, and information about whether or not a scar is present. In addition, the wound treatment expected information may include expected information on whether or not surgery and predictive information on whether or not a scar is generated.

The wound management device 220 applies the current patient's wound image data 520 to the machine learning model 510 to obtain wound depth information, wound treatment period information, surgery information, and scarring information. can do. The wound management device 220 additionally adds expected information on the presence or absence of surgery and expected information on whether scars are generated based on the acquired wound depth information, wound treatment period information, information on whether surgery, and information on whether scars are present. can be obtained

The wound management device 220 applies the current patient's wound image data 520 to the machine learning model 510 to obtain information on whether or not surgery and information on whether or not there is a scar, predictive information on whether or not surgery and The reason for additionally acquiring expected information on the presence or absence of scarring is to increase accuracy. Although the machine learning model 510 is rapidly developing, there are cases where the accuracy of the results is still poor. Therefore, the wound management device 220 is based on the information about the depth of the wound, the wound treatment period information, the operation information and the scarring information obtained from the machine learning model 510, and Accuracy of prediction can be increased by obtaining prediction information on whether scars have occurred.

The expected information on the presence or absence of surgery is information predicted by the wound management device 220 on whether surgery is required. For example, based on the current patient's wound image data and the machine learning model, the wound management apparatus 220 may determine whether surgery is required for the patient's wound. The user may determine whether surgery is necessary and quickly determine whether to visit the hospital based on the wound management device 220 .

In addition, the expected information on the occurrence of scars is information that predicts whether scars will occur after wound healing. The wound management apparatus 220 may determine whether or not a scar will occur after wound treatment, based on the current patient's wound image data and a machine learning model.

6 is a flowchart illustrating a process of acquiring expected information on the presence or absence of surgery according to an embodiment of the present disclosure.

Wound management device 220 is based on the information obtained by applying the machine learning model, information on whether the first scar is generated (a), the second information about the existence of scars (b), and the third information about the existence of scars Information (c) can be set.

According to an embodiment of the present disclosure, the wound management apparatus 220 performs a step 610 of setting information (a) on whether or not the first scar is generated based on the wound depth information obtained by applying the machine learning model. can be done When the wound depth information obtained by applying the machine learning model to the current patient's wound image data indicates a "first-degree image", the wound management device 220 sets the information (a) on whether or not the first scar has occurred to 0. You can follow the steps to set it up. In addition, the wound management device 220 performs the step of setting the information (a) on whether or not the first scar is generated to 1 when the acquired wound depth information indicates "second-degree burn" or "third-degree burn". can The wound management apparatus 220 may express "first-degree burns", "second-degree burns", and "third-degree burns" as indexes. For example, a “first-degree image” may be expressed as 0, a “second-degree image” may be expressed as 1, and a “third-degree image” may be expressed as 2.

According to another embodiment of the present disclosure, the wound management device 220 performs a step 610 of setting information (a) on whether or not the first scar is generated based on the wound depth information obtained by applying the machine learning model. can be done “Second-degree burns” can be divided into “deep-diseased” and “superficial”. "Superficial 2nd degree burn" means damage to the epidermis or the shallow dermal layer of the skin, and "deep dermal layer of the skin" means damage to the deep dermal layer of the skin. When the wound depth information obtained by applying a machine learning model to the current patient's wound image data indicates "first-degree burn" or "second-degree superficiality", the wound management device 220 provides information on whether or not a first scar has occurred The step of setting (a) to 0 may be performed. In addition, the wound management apparatus 220 sets the information (a) on whether or not the first scar has occurred to 1 when the acquired wound depth information indicates "second-degree burns", "third-degree burns" or "fourth-degree burns". You can follow the steps to set it up. The wound management apparatus 220 may express "first-degree burns", "second-degree burns superficially", "second-degree burns with superficial properties", and "third-degree burns" as indexes. For example, "first-degree burn" may be represented by 0, "second-degree image superficiality" may be represented by 1, "second-degree burn deepness" may be represented by 2, and "third-degree burn" may be represented by 3.

The wound management apparatus 220 may output the acquired wound depth information. The user may check the acquired wound depth information and determine how severe his or her wound is. In addition, the user may easily determine whether to visit the hospital based on the output of the wound management device 220 .

The wound management apparatus 220 may perform the step 620 of setting the information (b) on whether or not the second scar is generated based on the treatment period information of the wound obtained by applying the machine learning model. The wound management device 220 sets the information (b) on the presence or absence of the second scar to 0 when the treatment period information of the wound obtained by applying the machine learning model to the wound image data of the current patient is "7 days or less" steps can be performed. The wound management apparatus 220 may perform the step of setting the information (b) on whether or not the second scar is generated to 1 when the obtained wound treatment period information is “more than 7 days”. The wound management device 220 may express "7 days or less" and "more than 7 days" as an index. For example, "7 days or less" may be expressed as 0, and "more than 7 days" may be expressed as 1.

The wound management device 220 may output information about the treatment period of the obtained wound. The user may check the acquired wound treatment period information and estimate how long it will take to heal his or her wound. Accordingly, the user can prevent excessive treatment.

The wound management apparatus 220 may perform a step 630 of setting the information (c) on whether or not a third scar is generated based on the information on whether or not a scar is obtained by applying the machine learning model. The wound management device 220 sets the information (c) on whether or not the third scar is generated when the information on whether or not a scar obtained by applying a machine learning model to the wound image data of the current patient indicates “no scar remains” You can follow the steps to set it up. The wound management apparatus 220 may perform the step of setting the information (c) on whether or not the third scar is generated to 1 when the acquired information on whether or not a scar is present indicates “scar remains”. The wound management apparatus 220 may express "no scar left" and "scar left" as an index. For example, "no scar left" may be expressed as 0, and "scar remaining" may be expressed as 1.

The wound management device 220 may output information on whether or not a scar is obtained. Users can predict early whether a patient's wounds will leave scars. Accordingly, the user can quickly take necessary measures to minimize the scar on the patient's wound.

The wound management apparatus 220 may store the predetermined weights w1, w2, and w3 in a memory. As the importance of the information on whether or not the first scar is generated or the information on whether the third scar is generated is higher, the weight corresponding to the information on whether or not the first scar is generated or whether the third scar is generated may have a higher value. For example, w3 may be greater than w1 and w2. The fact that w3 is the largest indicates that the wound management device 220 highly evaluates the importance of information on whether or not the third scar is generated. Also, the wound management apparatus 220 may limit the sizes of w1, w2, and w3 for normalization of the weighted average. For example, the wound management apparatus 220 may set w1+w2+w3<=3.

The step 640 of calculating a weighted average ((w1*a+w2*b+w3*c)/3) of the information on whether the first scar is generated or the information about the third scar is performed ( 640 ). In addition, when the weighted average is greater than 0.5, the wound management apparatus 220 may perform the step 650 of determining the expected information on whether scars are generated as scars remaining.

The wound management apparatus 220 may output expected information on whether or not a scar is generated. Since the predicted information on the occurrence of scars is obtained based on the information on whether the first scar is generated or the predicted information on the presence or absence of the third scar, the reliability may be high. The wound management apparatus 220 may determine the need for a visit to the hospital based on the expected information on the presence or absence of scarring. The user may take necessary measures based on the expected information on the presence or absence of scarring.

When the weighted average ((w1*a+w2*b+w3*c)/3) of the information on whether or not the first scar is generated to the information on whether or not the third scar is generated is 0.5 or less, the wound management device 220 determines the scar A step 650 of determining that there is no scar left may be performed based on the expected information on the occurrence or not.

7 is a flowchart illustrating a process of acquiring expected information on whether scars occur according to an embodiment of the present disclosure.

The wound management apparatus 220 may set information (d) on whether or not the first operation is performed and information (e) on the presence or absence of the second operation based on information obtained by applying the machine learning model.

According to an embodiment of the present disclosure, the wound management device 220 sets the information (d) on the presence or absence of the first surgery based on the wound depth information obtained by applying a machine learning model to the wound image data of the current patient. step 710 may be performed. The wound management device 220 sets the information (d) on the presence or absence of the first operation to 0 when the wound depth information obtained by applying the machine learning model to the wound image data of the current patient indicates "first-degree image" steps can be performed. The wound management apparatus 220 may perform the step of setting the information (d) on the presence or absence of the first operation to 1/2 when the acquired depth information of the wound indicates a “second-degree image”. The wound management apparatus 220 may perform the step of setting the information (d) on the presence or absence of the first operation to 1 when the acquired wound depth information indicates a “third degree image”.

According to another embodiment of the present disclosure, the wound management apparatus 220 sets the information (d) on the presence or absence of the first operation based on the depth information of the wound obtained by applying the machine learning model to the wound image data of the current patient. step 710 may be performed. The wound management device 220 provides information on the presence or absence of the first operation ( The step of setting d) to 0 may be performed. The wound management apparatus 220 may perform the step of setting the information (d) on the presence or absence of the first operation to 1/2 when the acquired wound depth information indicates "second-degree burns". The wound management apparatus 220 may perform the step of setting the information (d) on the presence or absence of the first operation to 1 when the acquired wound depth information indicates a “third-degree image” or a “fourth-degree image”.

The wound management device 220 performs the step 720 of setting the information (e) on the presence or absence of the second operation based on the information on whether the operation is performed or not obtained by applying the machine learning model to the wound image data of the current patient. can The wound management device 220 sets the information (e) on the presence or absence of the second operation to 0 when the information on whether or not surgery obtained by applying the machine learning model to the wound image data of the current patient indicates "no surgery is required" You can follow the steps to set it up. The wound management apparatus 220 may perform the step of setting the information (e) on the presence or absence of the second operation to 1 when the obtained information on whether or not the operation is performed indicates "needs surgery". The wound management device 220 may express “no surgery required” and “need surgery” as an index. For example, “no surgery required” may be expressed as 0, and “needs surgery” may be expressed as 1.

The wound management device 220 may output information on whether the operation has been obtained. The user may predict the situation of the patient's wounds based on the obtained information on whether or not the operation has been performed. The user can visit the hospital so that the wound does not get worse and schedule the surgery quickly.

The wound management device 220 may store the predetermined weights w4 and w5 in a memory. The higher the importance of the information (d) on the presence or absence of the first operation and the information on the presence or absence of the second operation (e), the higher the information corresponds to the information on the presence or absence of the first operation (d) and the information on the presence or absence of the second operation (e) The weighted value may have a high value. w5 may be greater than w4. That is, the wound management device 220 may highly evaluate the information (e) on the presence or absence of the second operation. In addition, the wound management apparatus 220 may limit the sizes of w4 and w5 in order to normalize the value of the weighted average. For example, the wound management apparatus 220 may set w4+w5<=2.

The wound management apparatus 220 calculates a weighted average ((w4*d+w5*e)/2) of the information (d) on the presence or absence of the first operation and the information (e) on the presence or absence of the second operation (step 730) ) can be done. In addition, when the weighted average is greater than 0.5, the wound management apparatus 220 may perform a step 740 of determining the need for surgery based on expected information on the presence or absence of surgery.

When the weighted average ((w4*d+w5*e)/2) of the information (d) on the presence or absence of the first operation (d) and the information on the presence or absence of the second operation (e) is 0.5 or less, the wound management device 220 removes the scar A step 740 of determining that scarring does not occur may be performed based on the expected information on the presence or absence of occurrence.

The wound management apparatus 220 may output expected information on whether or not a scar is generated. Since the predicted information on the presence or absence of scarring is based on the information on the presence or absence of the first operation and the information on the presence or absence of the second operation, the reliability may be high. The wound management apparatus 220 may determine whether or not a visit is necessary based on the expected information on the presence or absence of scarring. In addition, the user can quickly take a necessary action based on the expected information on the presence or absence of scarring.

8 is a diagram illustrating a user terminal according to an embodiment of the present disclosure.

The user terminal 800 may include an input unit 210 and an output unit 230 . The user terminal 800 may include a wound management device 220 . Alternatively, the user terminal 800 may be a device different from the wound management device 220 . The user terminal 800 may transmit a signal received through the input unit 210 to the wound management apparatus 220 . In addition, the user terminal 800 may output a signal received from the wound management device 220 from the output unit 230 .

The user terminal 800 may display a wound image 811 . The user terminal 800 may display at least one of expected information 821 on whether or not surgery is performed and predictive information 822 on whether or not a scar is generated.

In addition, when the expected information 821 on the presence or absence of surgery indicates the need for surgery or the expected information 822 on the presence or absence of a scar indicates the occurrence of a scar, the user terminal 800 displays information 823 indicating that a visit is required. can do. The user can easily determine the patient's condition by using the user terminal 800 and quickly visit the hospital.

When it is determined that a visit is necessary, the wound management device 220 may acquire hospital information. For example, when the expected information on the presence or absence of surgery is set as the need for surgery, the wound management apparatus 220 may perform a step of acquiring hospital information. In addition, when the expected information on the presence or absence of scarring indicates scarring, the wound management apparatus 220 may perform a step of acquiring hospital information.

Specifically, the wound management device 220 may perform a step of requesting hospital information to the hospital server when a visit is required. The hospital server may include a processor and memory. The hospital server may store a database including information about the hospital and treatment information.

The hospital server may perform the step of transmitting hospital information to the wound management device 220 . The wound management device 220 may perform a step of outputting hospital information through an output unit.

The wound management apparatus 220 may acquire hospital information stored in the memory of the wound management apparatus 220 . In addition, the wound management device 220 may output the acquired hospital information through the output unit. In this case, the wound management device 220 may not require a hospital server for receiving hospital information.

The wound management device 220 may perform the step of outputting the obtained hospital information. Hospital information includes the contact information of the hospital, the location of the hospital, the distance between the hospital and the patient, the transportation to the hospital, the operating hours of the hospital, the information of the medical personnel belonging to the hospital, the opening date of the hospital, the number of re-uses by the patients in the hospital, and the rating of the hospital , it may include at least one of the current number of waiting people in the hospital, a parking space near the hospital, and a description of the hospital facility.

In addition, the hospital information may include information about a hospital that provided an image most similar to the image of the current patient's wound. The image similarity evaluation may be performed through various algorithms. In addition, the hospital information may include at least one of the time taken for the wound treatment by the hospital that provided the most similar image, the amount charged for the wound treatment, or the number of times the patient requested a visit.

Also, the wound management apparatus 220 may output an image of a change over time of a wound according to a treatment process of an image most similar to the image of the current patient's wound. The user can check the image of the change over time of the wound and determine whether his or her wound is being treated normally. Therefore, the wound management device 220 may reduce the anxiety that the user feels in the process of treating a patient's wound.

The patient can get information about the hospital that treated the wound most similar to his or her own. The more information a hospital provides about the treatment of wounds, the more likely it is that the hospital will be exposed to services. Through this, service providers can induce hospitals to autonomously provide treatment information for marketing.

The user can easily know information about the hospital for the treatment of wounds. In addition, patients can easily find out which hospitals have experience treating their wounds. In addition, patients can easily find out which hospital treated their wounds the fastest. In addition, the patient can easily know the hospital that treated his wounds at the lowest cost. In addition, patients can easily find out which hospital treated their wounds with fewer visits.

Wound management device 220 may perform the step of receiving the reservation information related to the hospital reservation from the user. The reservation information may include at least one of a patient's scheduled visit time and information on a medical person preferred by the patient.

Wound management device 220 may perform the step of transmitting information about the patient and reservation information to the hospital server. Wound management device 220 may perform the step of receiving the reservation confirmation signal from the hospital server.

So far, the case where the treatment information of the existing wound is information about the depth of the wound, the treatment period of the wound, whether the wound has been operated on or whether a scar is generated has been described. However, the present invention is not limited thereto. The existing wound treatment information may further include age information of the patient, gender information of the patient, wound healing period information, information on the number of visits to wound healing, information on the degree of scarring, and information on costs required to heal.

The wound management device 220 may generate a machine learning model by machine learning the correlation between image data of an existing wound and wound healing period information. The wound management apparatus 220 may generate a machine learning model by machine learning the correlation between image data of an existing wound and information on the number of visits up to wound healing. In addition, the wound management device 220 may acquire a machine learning model by learning the correlation between the image data of the existing wound and the degree of scarring. In addition, the wound management device 220 may acquire a machine learning model by learning the correlation between the image data of the existing wound and the cost information required to heal.

It may be difficult for the wound management device 220 to machine-learning wound healing period information, information on the number of visits to wound healing, information on the degree of scarring, and cost information until healing using only image data of an existing wound. Therefore, the wound management device 220 receives at least one of hospital identification information, medical personnel identification information, wound size information, information about the time that has passed since the wound occurred, information about the patient's age, and information about the patient's gender. You can acquire more to perform machine learning.

The wound management apparatus 220 may obtain wound treatment prediction information by applying the wound image data of the current patient to the generated machine learning model. The wound treatment expected information may include at least one of an expected wound healing period, information on the number of expected visits until wound healing, information on an expected degree of scarring, and information on estimated cost until healing.

The degree of effort required for the user to heal his or her wounds based on the expected wound healing period displayed on the wound management device 220, the expected number of visits to wound healing, information about the expected degree of scarring, and the estimated cost information until healing can be predicted Accordingly, there are cases in which the patient feels fear because of the uncertainty about wound treatment, and the wound management apparatus 220 of the present disclosure can reduce the patient's anxiety. In addition, the user can roughly determine the patient's condition without visiting a medical professional, thereby increasing convenience and preventing the patient from spending excessively on wound treatment.

In addition, since the wound management device 220 of the present disclosure judges the wound according to the same criteria, the possibility of receiving a different judgment depending on the medical personnel may be reduced. For example, in the case of a second-degree burn, insurance coverage may vary depending on whether it is superficial or deep, and there is a possibility that different medical personnel may make a different decision. Therefore, insurers and patients sometimes argue over whether it is deep or superficial. However, since the wound management device 220 of the present disclosure is determined based on the same criteria, the conflict between the insurance company and the patient can be reduced.

So far, various embodiments have been mainly looked at. Those of ordinary skill in the art to which the present invention pertains will understand that the present invention can be implemented in a modified form without departing from the essential characteristics of the present invention. Therefore, the disclosed embodiments are to be considered in an illustrative rather than a restrictive sense. The scope of the present invention is indicated in the claims rather than the foregoing description, and all differences within the scope equivalent thereto should be construed as being included in the present invention.

Meanwhile, the above-described embodiments of the present invention can be written as a program that can be executed on a computer, and can be implemented in a general-purpose digital computer that operates the program using a computer-readable recording medium. The computer-readable recording medium includes a storage medium such as a magnetic storage medium (eg, a ROM, a floppy disk, a hard disk, etc.) and an optically readable medium (eg, a CD-ROM, a DVD, etc.).

Claims (8)

acquiring image data of a plurality of existing wounds;
obtaining treatment information of an existing wound corresponding to the image data of the plurality of existing wounds;
acquiring a machine learning model by learning the correlation between the image data of the plurality of existing wounds and the treatment information of the existing wounds using a Convolution Neural Network (CNN);
Acquiring the wound image data of the current patient;
determining wound treatment prediction information by applying the machine learning model to the wound image data of the current patient; and
Including the step of outputting the wound treatment prediction information,
The treatment information of the existing wound includes all of the wound depth information, the wound treatment period information, the information on whether or not surgery, and the information on the scar,
The wound treatment prediction information includes expected information on the presence or absence of surgery and expected information on the presence or absence of scarring,
The wound depth information indicates one of "first-degree burns", "second-degree burns superficial", "second-degree burns deep", "third-degree burns" and "fourth-degree burns",
The treatment period information of the wound represents the number of days from the date of injury to the completion date of re-epithelialization,
The step of determining the wound treatment prediction information,
When the wound depth information predicted by applying the machine learning model to the wound image data of the current patient indicates the "first-degree burn" or the "second-degree image superficiality", information on whether or not a first scar has occurred (a ) is set to 0, and when the predicted wound depth information indicates the "second-degree burns", the "third-degree burns", or the "fourth-degree burns", information on whether or not the first scar has occurred (a ) to 1;
When the treatment period information of the predicted wound by applying the machine learning model to the wound image data of the current patient is "7 days" or less, the information (b) on the presence or absence of the second scar is set to 0, and the predicted wound setting the information (b) on whether or not the second scar is generated to 1 when the treatment period information is more than "7 days";
When the information on whether scars predicted by applying the machine learning model to the wound image data of the current patient indicates that no scars remain, the information (c) on whether or not a third scar occurs is set to 0, and the predicted scar setting the information (c) on whether or not a third scar is generated to 1 when the information on whether or not the scar remains;
Based on the predetermined weights (w1, w2, w3), a weighted average of the information on whether the first scar is generated or the information on the third scar is generated ((w1*a+w2*b+w3*c) /3) is greater than 0.5, determining the expected information on the presence or absence of scarring as a scar remaining; and
When the weighted average ((w1*a+w2*b+w3*c)/3) of the information on the presence or absence of the first scar formation to the information on the presence or absence of the third scar formation is 0.5 or less, the information on the presence or absence of the scar determining the predictive information to be non-scarred;
wherein w3 is greater than w1 and w2, and w1+w2+w3<=3.
delete The method of claim 1,
The step of determining the wound treatment prediction information,
When the wound depth information predicted by applying the machine learning model to the wound image data of the current patient indicates the "first-degree image" or the "second-degree image superficiality", information (d) on the presence or absence of the first operation set to 0, and when the predicted wound depth information indicates the "second-degree burn deepness", the information (d) on the presence or absence of the first operation is set to 1/2, and the predicted wound depth information is the " setting information (d) on whether or not the first operation is performed to 1 when indicating a third-degree burn or the "fourth-degree burn";
When the information on whether or not surgery predicted by applying the machine learning model to the wound image data of the current patient indicates "no need for surgery", the information (e) on the presence or absence of the second operation is set to 0, and the prediction setting information (e) on whether or not the second operation is performed to 1 when the information on whether or not the operation has been performed indicates "needs surgery";
Based on the predetermined weights (w4, w5), the weighted average ((w4*d+w5*e)/2 of the information (d) on the presence or absence of the first operation and the information (e) on the presence or absence of the second operation (e) ) is greater than 0.5, determining the expected information on the presence or absence of the operation as "requires surgery"; and
When the weighted average ((w4*d+w5*e)/2) of the information on the presence or absence of the first operation (d) and the information on the presence or absence of the second operation (e) is 0.5 or less, the prediction of the presence or absence of the operation determining the information as "no surgery required";
The w5 is greater than w4, w4+w5<=2, characterized in that the operating method of the wound care device.
4. The method of claim 3,
acquiring hospital information from a hospital server when the expected information on the scar occurrence is set to require surgery;
outputting the obtained hospital information;
Receiving reservation information related to a hospital reservation from a user;
transmitting patient information and reservation information to the hospital server; and
Further comprising the step of receiving a reservation confirmation signal from the hospital server,
The reservation information includes at least one of a patient's scheduled visit time and information on a medical person preferred by the patient,
The hospital information includes the contact information of the hospital, the location of the hospital, the distance between the hospital and the patient, transportation to the hospital, business hours of the hospital, information of medical personnel belonging to the hospital, the opening date of the hospital, the number of reuse of patients in the hospital, the hospital It further includes at least one of the rating of the hospital, the current number of waiting people in the hospital, a parking space near the hospital, and a description of the hospital facility,
The existing wound treatment information further includes patient age information, patient gender information, wound healing period information, information on the number of visits to wound healing, information on the degree of scarring, and cost information until healing,
The wound treatment estimate information includes an expected wound healing period, information on the expected number of visits until wound healing, information on the expected degree of scarring, and information on estimated costs until healing.
The wound care device comprises a processor and a memory;
The processor based on the instructions stored in the memory,
acquiring image data of a plurality of existing wounds;
obtaining treatment information of an existing wound corresponding to the image data of the plurality of existing wounds;
acquiring a machine learning model by learning the correlation between the image data of the plurality of existing wounds and the treatment information of the existing wounds using a Convolution Neural Network (CNN);
Acquiring the wound image data of the current patient;
determining wound treatment prediction information by applying the machine learning model to the wound image data of the current patient; and
performing the step of outputting the wound treatment prediction information,
The treatment information of the existing wound includes all of the wound depth information, the wound treatment period information, the information on whether or not surgery, and the information on the scar,
The wound treatment prediction information includes expected information on the presence or absence of surgery and expected information on the presence or absence of scarring,
The wound depth information indicates one of "first-degree burns", "second-degree burns superficial", "second-degree burns deep", "third-degree burns" and "fourth-degree burns",
The treatment period information of the wound represents the number of days from the date of injury to the completion date of re-epithelialization,
The processor is configured to determine the wound treatment prediction information,
When the wound depth information predicted by applying the machine learning model to the wound image data of the current patient indicates the "first-degree burn" or the "second-degree image superficiality", information on whether or not a first scar has occurred (a ) is set to 0, and when the predicted wound depth information indicates the "second-degree burns", the "third-degree burns" or the "fourth-degree burns", information on whether or not the first scar has occurred setting (a) to 1;
When the treatment period information of the predicted wound by applying the machine learning model to the wound image data of the current patient is "7 days" or less, the information (b) on the presence or absence of the second scar is set to 0, and the predicted wound setting the information (b) on whether or not the second scar is generated to 1 when the treatment period information is more than "7 days";
When the information on whether scars predicted by applying the machine learning model to the wound image data of the current patient indicates that no scars remain, the information (c) on whether or not a third scar occurs is set to 0, and the predicted scar setting information (c) on whether or not a third scar is generated to 1 when the information on whether or not there is a scar remains;
Based on the predetermined weights (w1, w2, w3), the weighted average of the information on whether the first scar is generated or the information on the third scar is generated ((w1*a+w2*b+w3*c) /3) is greater than 0.5, determining the expected information on the presence or absence of scarring as a scar remaining; and
When the weighted average ((w1*a+w2*b+w3*c)/3) of the information on the presence or absence of the first scar formation to the information on the presence or absence of the third scar formation is 0.5 or less, the information on the presence or absence of the scar performing the steps of determining the expected information to be non-scarring;
wherein w3 is greater than w1 and w2, and w1+w2+w3<=3.
delete 6. The method of claim 5,
The processor is configured to determine the wound treatment prediction information,
When the wound depth information predicted by applying the machine learning model to the wound image data of the current patient indicates the "first-degree image" or the "second-degree image superficiality", information (d) on the presence or absence of the first operation set to 0, and when the predicted wound depth information indicates the "second-degree burn deepness", the information (d) on the presence or absence of the first operation is set to 1/2, and the predicted wound depth information is the " setting information (d) on whether or not the first operation is performed to 1 when indicating a third-degree burn or the "fourth-degree burn";
When the information on whether or not surgery predicted by applying the machine learning model to the wound image data of the current patient indicates "no need for surgery", the information (e) on the presence or absence of the second operation is set to 0, and the prediction setting information (e) on whether or not the second operation is performed to 1 when the information on whether or not the operation has been performed indicates "needs surgery";
Based on the predetermined weights (w4, w5), the weighted average ((w4*d+w5*e)/2 of the information (d) on the presence or absence of the first operation and the information (e) on the presence or absence of the second operation (e) ) is greater than 0.5, determining the expected information on the presence or absence of the operation as "requires surgery"; and
When the weighted average ((w4*d+w5*e)/2) of the information on the presence or absence of the first operation (d) and the information on the presence or absence of the second operation (e) is 0.5 or less, the prediction of the presence or absence of the operation performing the steps to determine the information as "no surgery required";
The w5 is greater than w4, and a wound care device, characterized in that w4+w5<=2.
8. The method of claim 7,
The processor based on the instructions stored in the memory,
acquiring hospital information from a hospital server when the expected information on the scar occurrence is set to require surgery;
outputting the obtained hospital information;
Receiving reservation information related to a hospital reservation from a user;
transmitting patient information and reservation information to the hospital server; and
Further performing the step of receiving a reservation confirmation signal from the hospital server,
The reservation information includes at least one of a patient's scheduled visit time and information on a medical person preferred by the patient,
The hospital information includes the contact information of the hospital, the location of the hospital, the distance between the hospital and the patient, transportation to the hospital, business hours of the hospital, information of medical personnel belonging to the hospital, the opening date of the hospital, the number of reuse of patients in the hospital, the hospital It further includes at least one of the rating of the hospital, the current number of waiting people in the hospital, a parking space near the hospital, and a description of the hospital facility,
The existing wound treatment information further includes patient age information, patient gender information, wound healing period information, information on the number of visits to wound healing, information on the degree of scarring, and cost information until healing,
The wound treatment estimate information is wound care apparatus, characterized in that it further comprises an expected wound healing period, information on the expected number of visits until wound healing, information on the expected degree of scarring, and information on estimated cost until healing.

Images