KR20240004197A - Apparatus for refining liguid and diagnosis system including the same - Google Patents

Abstract

A liquid purification device is disclosed. The liquid purification device includes a substrate, a loading portion formed on top of the substrate and containing a first liquid, and a second liquid in which the concentration of at least one substance is reduced by reducing the concentration of at least one component contained in the first liquid. A filter unit that obtains, mixes the second liquid with the reactive material for detecting the target material, so that among the plurality of materials contained in the second liquid, the first material that has undergone a preset reaction with the reactive material and the first material that has not undergone a preset reaction with the reactive material It includes a reaction unit for obtaining a third liquid containing a second substance, and a separation unit for separating the first substance and the second substance.

Description

Liquid purification device and diagnostic system including the same {APPARATUS FOR REFINING LIGUID AND DIAGNOSIS SYSTEM INCLUDING THE SAME}

The embodiment relates to a liquid purification device.

The embodiment relates to a medical diagnostic device.

The embodiment relates to a medical diagnosis system.

Research is being actively conducted on technologies for diagnosing the health status or presence of disease of a subject based on liquid collected from the subject (e.g., blood, urine, etc.). For example, the technology can obtain diagnostic results for a subject by analyzing spectral information about a liquid. Typically, the above technology obtains diagnostic results based on spectral information about blood in which an antigen-antibody reaction has occurred.

However, in addition to antigens, various components, including blood cell components, exist in blood, which causes spectral information to contain errors. Therefore, in order to obtain accurate diagnostic results, technology is needed to prevent or compensate for errors in spectrum information.

One technical problem to be solved by the present invention is to provide a liquid purification device that can improve the accuracy of diagnostic results for a subject.

Another technical problem to be solved by the present invention is to provide a diagnostic device that can compensate for errors in spectral information about a target material.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description below.

According to an exemplary embodiment of the present disclosure for solving the above-described technical problem, a liquid purification device includes: a substrate; a loading portion formed on an upper portion of the substrate and containing a first liquid; a filter unit that reduces the concentration of at least one component contained in the first liquid to obtain a second liquid in which the concentration of the at least one substance is reduced; The second liquid is mixed with a reactive material for detecting the target material, so that among the plurality of materials included in the second liquid, a first material that undergoes a predetermined reaction with the reactive material and a first material that does not undergo a predetermined reaction with the reactive material a reaction unit that obtains a third liquid containing a second substance that is not present; and a separation unit for separating the first material and the second material.

According to an exemplary embodiment of the present disclosure for solving the above-described technical problem, in a diagnostic system including a liquid purification device, a liquid information acquisition device, and a diagnostic device, the liquid purification device includes: a substrate; a loading portion formed on an upper portion of the substrate and receiving a first liquid collected from a subject; a filter unit that reduces the concentration of at least one component contained in the first liquid to obtain a second liquid in which the concentration of the at least one substance is reduced; The second liquid is mixed with a reactive material for detecting the target material, so that among the plurality of materials included in the second liquid, a first material that undergoes a predetermined reaction with the reactive material and a first material that does not undergo a predetermined reaction with the reactive material a reaction unit that obtains a third liquid containing a second substance that is not present; and a separation unit for separating the first material and the second material, wherein the liquid information acquisition device acquires a Raman signal for the first material by irradiating light to the first material, The diagnostic device may be provided with a diagnostic system that obtains a diagnostic result for the subject based on the Raman signal.

According to an exemplary embodiment of the present disclosure for solving the above-described technical problem, a diagnostic device includes: a memory storing at least one instruction; and a processor, wherein the processor, by executing the at least one instruction, collects information about a first spectrum corresponding to a first liquid collected from a subject and containing a first target material, and the first liquid. As a first reactive material for detection of the first target material is added, information on a second spectrum corresponding to a second liquid containing the first target material combined with the first reactive material is obtained, Obtaining first concentration information of the first target material based on the information about the first spectrum and the information about the second spectrum, and obtaining diagnostic information about the subject based on the first concentration information. A diagnostic device may be provided.

According to an exemplary embodiment of the present disclosure for solving the above-described technical problem, in a method for controlling a diagnostic device, the first spectrum corresponding to the first liquid collected from the subject and containing the first target material is provided. information, and a second spectrum corresponding to a second liquid containing the first target material bound to the first reactive material as the first reactive material for detection of the first target material is added to the first liquid. Obtaining information about; Obtaining first concentration information of the first target material based on information about the first spectrum and information about the second spectrum; A control method including a step of acquiring diagnostic information about the subject based on the first concentration information may be provided.

The means for solving the problem of the present disclosure are not limited to the above-mentioned solution means, and the solution methods not mentioned can be clearly understood by those skilled in the art from this specification and the attached drawings. You will be able to.

According to various embodiments of the present disclosure as described above, errors in spectral information about liquid collected from a subject can be reduced. A diagnostic device can compensate for errors in spectral information. Accordingly, the accuracy of diagnostic results for the subject can be improved.

In addition, effects that can be obtained or expected due to the embodiments of the present disclosure will be disclosed directly or implicitly in the detailed description of the embodiments of the present disclosure. For example, various effects expected according to embodiments of the present disclosure will be disclosed in the detailed description to be described later.

Other aspects, advantages and salient features of the present disclosure will become apparent to those skilled in the art from the following detailed description, taken in conjunction with the accompanying drawings, which disclose various embodiments of the present invention.

Aspects, features and advantages of specific embodiments of the present disclosure will become clearer through the following description with reference to the accompanying drawings.
1 is a block diagram showing the configuration of a diagnostic system according to an embodiment of the present disclosure.
Figure 2 is a block diagram showing the configuration of a liquid purification device according to an embodiment of the present disclosure.
Figure 3 is a view from the top of a liquid purification device according to an embodiment of the present disclosure.
Figure 4 is a view of a liquid purification device according to an embodiment of the present disclosure viewed from below.
Figure 5 is a view from the top of a liquid purification device according to an embodiment of the present disclosure.
Figure 6 is a diagram showing a filter unit according to an embodiment of the present disclosure.
Figure 7 is a diagram showing a separation unit according to an embodiment of the present disclosure.
FIG. 8 is a diagram for explaining a method of acquiring liquid information by a liquid information acquisition device according to an embodiment of the present disclosure.
Figure 9 is a block diagram showing the configuration of a diagnostic device according to an embodiment of the present disclosure.
Figure 10 is a diagram showing information on a spectrum according to an embodiment of the present disclosure.
Figure 11 is a diagram for explaining a method of obtaining concentration information according to the first embodiment of the present disclosure.
FIG. 12 is a diagram for explaining a method of learning a first neural network model according to an embodiment of the present disclosure.
Figure 13 is a diagram for explaining a method of obtaining concentration information according to a second embodiment of the present disclosure.
Figure 14 is a diagram for explaining a method of obtaining concentration information according to a third embodiment of the present disclosure.
Figure 15 is a diagram for explaining a method of obtaining concentration information according to a fourth embodiment of the present disclosure.
Figure 16 is a diagram for explaining a method of obtaining concentration information according to an embodiment of the present disclosure.
Figure 17 is a diagram for explaining a method of obtaining concentration information according to an embodiment of the present disclosure.
Figure 18 is a diagram for explaining a method of obtaining concentration information according to an embodiment of the present disclosure.
Figure 19 is a flowchart showing a control method of a diagnostic device according to an embodiment of the present disclosure.

Terms used in this specification will be briefly described, and the present disclosure will be described in detail.

The terms used in the embodiments of the present disclosure have selected general terms that are currently widely used as much as possible while considering the functions in the present disclosure, but this may vary depending on the intention or precedent of a technician working in the art, the emergence of new technology, etc. . In addition, in certain cases, there are terms arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the description part of the relevant disclosure. Therefore, the terms used in this disclosure should be defined based on the meaning of the term and the overall content of this disclosure, rather than simply the name of the term.

Embodiments of the present disclosure may be subject to various changes and may have various embodiments, and specific embodiments will be illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the scope to specific embodiments, and should be understood to include all transformations, equivalents, and substitutes included in the disclosed spirit and technical scope. In describing the embodiments, if it is determined that a detailed description of related known technology may obscure the point, the detailed description will be omitted.

Terms such as first, second, etc. may be used to describe various components, but the components should not be limited by the terms. Terms are used only to distinguish one component from another.

Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, terms such as “comprise” or “consist of” are intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but are intended to indicate the presence of one or more other It should be understood that this does not exclude in advance the presence or addition of features, numbers, steps, operations, components, parts, or combinations thereof.

Below, with reference to the attached drawings, embodiments of the present disclosure will be described in detail so that those skilled in the art can easily practice them. However, the present disclosure may be implemented in many different forms and is not limited to the embodiments described herein. In order to clearly explain the present disclosure in the drawings, parts that are not related to the description are omitted, and similar parts are given similar reference numerals throughout the specification.

According to an exemplary embodiment of the present disclosure, a liquid purification device comprising: a substrate; a loading portion formed on an upper portion of the substrate and containing a first liquid; a filter unit that reduces the concentration of at least one component contained in the first liquid to obtain a second liquid in which the concentration of the at least one substance is reduced; The second liquid is mixed with a reactive material for detecting the target material, so that among the plurality of materials included in the second liquid, a first material that undergoes a predetermined reaction with the reactive material and a first material that does not undergo a predetermined reaction with the reactive material a reaction unit that obtains a third liquid containing a second substance that is not present; and a separation unit for separating the first material and the second material.

The first liquid may include blood in an undiluted state, and the at least one component may include a blood cell component.

The predetermined reaction may include an antigen-antibody reaction.

The liquid separation device may further include an anticoagulation unit that adds an anticoagulant to the first liquid.

The reaction unit may include a reactive material storage unit that stores the reactive material, and a zigzag-shaped mixing channel to increase the mixing ratio of the second liquid and the reactive material.

The reactive material may be bound to metal nanoparticles bound to Raman active molecules.

The liquid purification device includes: a first chamber formed on the upper part of the substrate, connected to the loading unit to store the first liquid; a second chamber formed on the upper part of the substrate and connected to the filter unit to store the second liquid; and a third chamber formed on the upper part of the substrate, connected to the separation part, and storing the first material.

The liquid purification device further includes a concentration unit that performs concentration processing on the third liquid, and the concentration processing may include at least one of drying, heating, and baking.

The liquid purification device further includes a pumping unit for moving the first liquid, the second liquid, and the third liquid, and the pumping unit includes at least one of a pneumatic pump, a vibration pump, a mechanical pump, and a capillary pump. may include.

The reaction unit includes a first reactant storage unit that stores a first reactant for detecting the first target, and a first mixing channel for increasing the mixing ratio of the second liquid and the first reactant. A second reaction comprising a reaction unit, a second reaction material storage unit for storing a second reaction material for detecting a second target, and a second mixing channel for increasing the mixing ratio of the second liquid and the second reaction material. It may include a unit, a first channel for transferring the second liquid to the first reaction unit, and a second channel for transferring the second liquid to the second reaction unit.

The first channel and the second channel may have a structure branched from the output terminal of the filter unit.

The separation unit is connected to the output terminal of the first reaction unit and serves as a first separation unit for separating a third substance that has undergone an antigen-antibody reaction with the first reaction substance and other substances among a plurality of substances included in the third liquid. and a second separation unit connected to the output terminal of the second reaction unit and configured to separate a fourth substance that has undergone an antigen-antibody reaction with the second reaction substance and other substances among the plurality of substances contained in the third liquid. It can be included.

The filter unit may include a Lateral Cavity Acoustic Transducer (LCAT) that protrudes to the outside of the filter channel through which the first liquid flows.

The separation unit separates the first material and the second material according to molecular weight based on sound waves, and includes a first outlet channel for moving the first material and a second outlet channel for moving the second material. It can be included.

According to an exemplary embodiment of the present disclosure, in a diagnostic system including a liquid purification device, a liquid information acquisition device, and a diagnostic device, the liquid purification device includes: a substrate; a loading portion formed on an upper portion of the substrate and receiving a first liquid collected from a subject; a filter unit that reduces the concentration of at least one component contained in the first liquid to obtain a second liquid in which the concentration of the at least one substance is reduced; The second liquid is mixed with a reactive material for detecting the target material, so that among the plurality of materials included in the second liquid, a first material that undergoes a predetermined reaction with the reactive material and a first material that does not undergo a predetermined reaction with the reactive material a reaction unit that obtains a third liquid containing a second substance that is not present; and a separation unit for separating the first material and the second material, wherein the liquid information acquisition device acquires a Raman signal for the first material by irradiating light to the first material, The diagnostic device may be provided with a diagnostic system that obtains a diagnostic result for the subject based on the Raman signal.

According to an exemplary embodiment of the present disclosure, a diagnostic device includes: a memory storing at least one instruction; and a processor, wherein the processor, by executing the at least one instruction, collects information about a first spectrum corresponding to a first liquid collected from a subject and containing a first target material, and the first liquid. As a first reactive material for detection of the first target material is added, information on a second spectrum corresponding to a second liquid containing the first target material combined with the first reactive material is obtained, Obtaining first concentration information of the first target material based on the information about the first spectrum and the information about the second spectrum, and obtaining diagnostic information about the subject based on the first concentration information. A diagnostic device may be provided.

The first liquid may not contain the first reactive material.

The information about the second spectrum includes a peak value of the second spectrum, and the processor, by executing the at least one instruction, matches the peak value of the spectrum and the concentration information based on a table pre-stored in the memory. Concentration information corresponding to the peak value of the second spectrum is obtained, information on the first spectrum is input into a first neural network model to obtain a coefficient for correcting the concentration information, and based on the coefficient, the The first concentration information can be obtained by correcting the concentration information.

By executing the at least one instruction, the processor may acquire the first concentration information by inputting information about the first spectrum and information about the second spectrum into a second neural network model.

The processor, by executing the at least one instruction, inputs information about the second spectrum into a second neural network model to obtain a feature vector, and inputs the information about the first spectrum and the feature vector into a third neural network model. The first concentration information can be obtained by entering .

By executing the at least one instruction, the processor inputs information about the first spectrum into a second neural network model to obtain a first feature vector, and inputs information about the second spectrum into the second neural network model. A second feature vector can be obtained by inputting the first feature vector and the second feature vector into a fourth neural network model to obtain the first concentration information.

By executing the at least one instruction, the processor may obtain the diagnostic information by inputting information about the first spectrum and the first concentration information into a fifth neural network model.

The processor, by executing the at least one instruction, detects the second target material combined with the second reactive material as the second reactive material for detection of the second target material contained in the first liquid is added. Obtaining information on a third spectrum corresponding to a third liquid containing, and obtaining second concentration information of the second target material based on the information on the first spectrum and the information on the third spectrum, The diagnostic information may be obtained by inputting information about the first spectrum, the first concentration information, and the second concentration information into the fifth neural network model.

By executing the at least one instruction, the processor may obtain the diagnostic information by inputting information about the first spectrum and information about the second spectrum into a sixth neural network model.

According to an exemplary embodiment of the present disclosure, in a method for controlling a diagnostic device, information on a first spectrum corresponding to a first liquid collected from a subject and containing a first target material, and As a first reactive material for detection of the first target material is added, obtaining information on a second spectrum corresponding to a second liquid containing the first target material bound to the first reactive material; Obtaining first concentration information of the first target material based on information about the first spectrum and information about the second spectrum; A control method including a step of acquiring diagnostic information about the subject based on the first concentration information may be provided.

The step of acquiring the first concentration information includes obtaining concentration information corresponding to the peak value of the second spectrum based on a table in which the peak value of the spectrum and the concentration information are matched and stored in advance in the memory, the first It may include inputting information about the spectrum into a first neural network model to obtain a coefficient for correcting the concentration information, and obtaining the first concentration information by correcting the concentration information based on the coefficient. there is.

In the step of acquiring the first concentration information, the first concentration information may be obtained by inputting information about the first spectrum and information about the second spectrum into a second neural network model.

Obtaining the first concentration information includes acquiring a feature vector by inputting information about the second spectrum into a second neural network model, and applying the information about the first spectrum and the feature vector to a third neural network model. It may include obtaining the first concentration information by inputting it into .

Obtaining the first concentration information includes obtaining a first feature vector by inputting information about the first spectrum into a second neural network model, inputting information about the second spectrum into the second neural network model. It may include obtaining a second feature vector, and acquiring the first concentration information by inputting the first feature vector and the second feature vector into a fourth neural network model.

In the step of acquiring the diagnostic information, the diagnostic information may be obtained by inputting information about the first spectrum and the first concentration information into a fifth neural network model.

The control method corresponds to a third liquid containing the second target material bound to the second reactant as a second reactive material for detection of the second target material contained in the first liquid is added. Obtaining information about a third spectrum; and acquiring second concentration information of the second target material based on the information about the first spectrum and the information about the third spectrum, wherein the step of obtaining diagnostic information includes: The diagnostic information can be obtained by inputting information about, the first concentration information, and the second concentration information into the fifth neural network model.

In the step of acquiring the diagnostic information, the diagnostic information may be obtained by inputting information about the first spectrum and information about the second spectrum into a sixth neural network model.

1 is a block diagram showing the configuration of a diagnostic system according to an embodiment of the present disclosure.

Referring to FIG. 1 , the diagnostic system 1000 may include a liquid purification device 100, a liquid information acquisition device 200, and a diagnostic device 300.

The liquid purification device 100 is a device for separating liquid. For example, the liquid purification device 100 can separate blood cell components and plasma components of blood. Alternatively, the liquid purification device 100 may separate first blood to which the first reactive material has been added and second blood to which the second reactive material has been added. The liquid may include at least one of the subject's urine, saliva, semen, sweat, tears, and cerebrospinal fluid. The subject may be a human or an animal.

The liquid purification device 100 may be implemented as a biochip.

The liquid information acquisition device 200 is a device for acquiring liquid information, which is information about the liquid. Liquid information may include spectrum information corresponding to the liquid. Spectral information may represent intensity by wavelength (or frequency). Spectral information (also referred to as information about the spectrum) may include Raman signals.

The liquid information acquisition device 200 may acquire spectral information based on Raman scattering that occurs when light passes through a certain medium. For example, the liquid information acquisition device 200 may irradiate a laser beam to the liquid obtained by the liquid purification device 100 and obtain a laser beam scattered from fine particles in the liquid. The liquid purification device 100 may acquire spectral information corresponding to the liquid based on the scattered laser beam. Spectral information, that is, the intensity of the wavelength and wavelength band of the Raman signal, may vary depending on the components of the microparticles.

The liquid information acquisition device 200 may include intensity corresponding only to a specific frequency among the spectrum information. For example, when a Raman active molecule is bound to a fine particle, the liquid information acquisition device 200 may acquire the peak value of the Raman signal at a frequency corresponding to the Raman active molecule.

The liquid information acquisition device 200 may include a light source for irradiating light to the liquid. The light source may emit a laser beam to induce Raman scattering from fine particles in the liquid.

The liquid information acquisition device 200 may include an optical system. The optical system may include a component for transmitting light irradiated from a light source to the liquid and sensing light scattered from fine particles in the liquid. For example, the optical system may include at least one of a filter, mirror, lens, slit, grating, and sensor. The sensor can sense light scattered from fine particles in a liquid.

The liquid information acquisition device 200 may be implemented as a spectrometer.

The diagnostic device 300 is configured to obtain various analysis data related to liquid. The diagnostic device 300 may be a medical diagnostic device. For example, the diagnostic device 300 may obtain identification information and concentration information about a plurality of substances constituting the liquid obtained by the liquid information acquisition device 200. Additionally, the diagnostic device 300 may obtain the concentration ratio of a plurality of substances contained in the liquid.

The diagnostic device 300 can obtain various diagnostic results. For example, the diagnostic device 300 may obtain a diagnostic result including the health condition or presence or absence of a disease corresponding to the subject. The diagnosis result may include identification information about the disease that the subject is expected to have and probability information that the subject has the disease.

The diagnostic device 300 may be implemented as a server or a user terminal device.

Figure 2 is a block diagram showing the configuration of a liquid purification device according to an embodiment of the present disclosure.

Referring to FIG. 2, the liquid purification device 100 includes a loading unit 110, an anticoagulation unit 120, a filter unit 130, a reaction unit 140, a concentrating unit 160, and a separation unit 150. It can be included.

The loading unit 110 may include an injection port through which liquid is injected and a channel for moving the liquid. The loading unit 110 may receive liquid through an injection port and transfer the received liquid to other components of the liquid purification device 100. For example, the loading unit 110 may receive the first liquid and transfer some of the first liquid to the anticoagulation unit 120 . Alternatively, the loading unit 110 may deliver some of the first liquid to the chamber.

The anticoagulation unit 120 is configured to perform anticoagulation treatment on liquid. The anticoagulation unit 120 may include a chamber in which an anticoagulant is stored. For example, the anticoagulation unit 120 may be located on a path along which the first liquid received by the loading unit 110 moves to the filter unit 130. The anticoagulation unit 120 may include an inlet that receives the first liquid and an outlet that outputs the anticoagulated first liquid. Alternatively, the anticoagulant unit 120 may include a pipe for injecting an anticoagulant into a channel through which the first liquid flows to the filter unit 130.

The filter unit 130 can accommodate anticoagulant treated liquid. The filter unit 130 may reduce the concentration of at least one component contained in the received liquid. For example, at least one component may include a blood cell component. Blood cell components may include red blood cells, white blood cells, and platelets.

The filter unit 130 may include a filter channel through which liquid flows and a fluid structure for filtering at least one component contained in the liquid. The filter channel may accommodate the first liquid that has been anticoagulated by the anticoagulation unit 120. The fluid structure protrudes outward from the filter channel and may include protrusions containing air pockets. For example, the fluid structure may include a Lateral Cavity Acoustic Transducer (LCAT).

The filter unit 130 may filter at least one component contained in the liquid based on vibration transmitted to the filter channel and air bladder. When vibration is transmitted to the filter channel and the air bladder, a vortex of liquid may be generated at the interface of the filter channel and the air bladder. The vortices in the liquid may capture at least one component contained in the liquid passing through the filter channel. Accordingly, the filter unit 130 can obtain a liquid with a reduced concentration of at least one component. For example, the filter unit 130 may receive the anticoagulated first liquid and filter blood cell components included in the anticoagulated first liquid. Accordingly, the filter unit 130 can obtain the second liquid with a reduced concentration of blood cell components. The second liquid may be plasma or serum.

The reaction unit 140 may induce a preset reaction between the reaction material for detecting the target material and the target material contained in the liquid. The target substance may include an antigen. Reactive substances may include antibodies. The reactive material may be bound to a metal (eg, gold) nanoparticle to which a Raman active molecule (or Raman reporter) is bound. A predetermined reaction refers to a chemical reaction, and an example may include an antigen-antibody reaction. Multiple metal nanoparticles can bind to antigens through antibodies. Accordingly, one metal nanoparticle lump in which a plurality of metal nanoparticles are combined with the antigen at the center may be formed.

The reaction unit 140 may obtain the third liquid by reacting the second liquid and the reaction material. The second liquid may include a reactive material, a first material in which a predetermined reaction occurs, and a second material in which a predetermined reaction does not occur. The third liquid may include a first material that has undergone a predetermined reaction with the reactive material and a second material that has not undergone a predetermined reaction with the reactive material.

The reaction unit 140 may include a reaction material storage unit that stores a reaction material. The reaction unit 140 may include a mixing channel to increase the mixing ratio of the second liquid and the reaction material. For example, the mixing channel may have a zigzag shape.

The separation unit 150 is configured to separate substances contained in the liquid. For example, the separator 150 may separate the first material and the second material included in the third liquid. The separation unit 150 may separate the first material and the second material based on various methods. As an example, the separation unit 150 may separate the first material and the second material according to molecular weight based on surface acoustic waves (SAW). The separation unit 150 may include a surface acoustic wave filter.

The separation unit 150 may include a channel for moving a plurality of separated materials. For example, the separation unit 150 may include a first outlet channel for moving the first material and a second outlet channel for moving the second material.

The enrichment unit 160 is configured to perform concentration processing on the liquid to increase the concentration of the target substance contained in the liquid. Concentration treatment may include at least one of drying, heating, and baking. Alternatively, the concentrator 160 may include a filter (eg, a membrane filter) or a pipe that allows only substances smaller than a preset size to pass. At this time, the concentrator 160 may allow materials smaller in size than the first metal nanoparticles in a state bound to the target material to pass through the reaction material, but may not pass the first metal nanoparticles. Accordingly, the concentrator 160 can obtain liquid with an increased concentration of first metal nanoparticles. That is, the concentrator 160 can obtain liquid with an increased concentration of the target substance.

The concentrator 160 may obtain the fourth liquid containing the first substance separated by the separator 150. The fourth liquid may include first metal nanoparticles bound to the first material through an antibody. The fourth liquid may include a third material (eg, protein) whose molecular weight is greater than that of the first metal nanoparticles or a second material that is not accurately separated by the separator 150. Alternatively, the fourth liquid may include second metal nanoparticles that are not combined with the first material. The concentrator 160 allows at least a portion of the remaining materials except the first metal nanoparticles to pass through, but does not pass the first metal nanoparticles, thereby obtaining a fifth liquid with an improved concentration of the first metal nanoparticles. Accordingly, the enrichment unit 160 can obtain a fifth liquid with an improved concentration of the first substance compared to the fourth liquid.

Figure 3 is a view from the top of a liquid purification device according to an embodiment of the present disclosure. Figure 4 is a view of a liquid purification device according to an embodiment of the present disclosure viewed from below.

Referring to FIG. 3, the liquid purification device 100 includes a substrate 10, a loading unit 110, an anticoagulation unit 120, a filter unit 130, a reaction unit 140, a separation unit 150, and a concentration unit. It may include a portion 160 and chambers 31, 32, 33, 34, and 35.

The loading part 110 may be formed on the upper part of the substrate 10 . The loading unit 110 can accommodate the first liquid. The loading unit 110 may deliver some of the first liquid to the first chamber 31. The first chamber 31 may store the delivered first liquid.

The loading unit 110 may deliver the first liquid to the anticoagulation unit 120. The anticoagulation unit 120 may perform anticoagulation treatment on the first liquid. The anticoagulation unit 120 may deliver the anticoagulated first liquid to the filter unit 130.

The filter unit 130 includes a filter channel 131 through which the first liquid flows and a fluid structure 132 that protrudes outward from the filter channel 131 and includes an air bladder, the filter channel 131, and the fluid structure 132. It may include a first vibration generator 133 that provides sound waves.

The filter unit 130 may filter at least one first component (eg, blood cell component) included in the first liquid based on the sound wave generated by the first vibration generator 133. The filter channel 131 and the fluid structure 132 may vibrate due to the sound waves generated by the first vibration generator 133. As the filter channel 131 and the fluid structure 132 vibrate, a vortex of the first liquid may be generated at the interface between the filter channel 131 and the air bladder. At least one first component may be trapped in the vortex of the produced first liquid. Accordingly, the filter unit 130 may obtain a second liquid in which the concentration of at least one first component is reduced compared to the first liquid.

The filter unit 130 may move the first liquid and the second liquid based on the air pocket included in the fluid structure 132. The air bladder may repeat compression and expansion based on the sound waves generated by the first vibration generator 133. The bladder can pump the first liquid and the second liquid. The filter unit 130 may output the second liquid. The fluid structure 132 and the first vibration generator 133 may form a Lateral Cavity Acoustic Transducer (LCAT).

Referring to FIG. 4 , the first vibration generator 133 may be located below the substrate 10 . The first vibration generator 133 may include a plurality of electrodes 1331 and 1332. The plurality of electrodes 1331 and 1332 may be arranged to face each other. The plurality of electrodes 1331 and 1332 may be piezoelectric electrodes.

Referring again to FIG. 3 , the filter unit 130 may transfer some of the second liquid to the second chamber 32 . The second liquid may be blood from which blood cells have been removed from the first liquid.

The filter unit 130 may deliver the second liquid to the reaction unit 140. The reaction unit 140 may include a reactive material storage unit 141 that stores a reactive material bound to metal nanoparticles. Raman active molecules may be bound to metal nanoparticles.

The second liquid may be mixed with the reactive material while passing through the reactive material storage unit 141. An antigen-antibody reaction may occur between the target material (or first material) contained in the second liquid and the reactive material. In this process, a metal nanoparticle mass in which at least one metal nanoparticle is clustered around an antigen may be formed.

The reaction unit 140 may include a zigzag-shaped mixing channel 142 to increase the mixing ratio of the second liquid and the reaction material. While passing through the mixing channel 142, an antigen-antibody reaction may occur between the target material and the reactive material. Accordingly, the number of metal nanoparticle lumps may increase. Although not shown, the reaction unit 140 may include a Lateral Cavity Acoustic Transducer (LCAT) to increase the mixing ratio of the second liquid and the reaction material.

The reaction unit 140 may obtain the third liquid based on the second liquid. The third liquid may include a first material that reacts with the reactive material to form a metal nanoparticle clump, and a plurality of second materials that do not react with the reactive material to form a metal nanoparticle clump. The reaction unit 140 may deliver the third liquid to the separation unit 150.

The separation unit 150 may include a second vibration generator 151 that generates sound waves, a first outlet channel 152, and a second outlet channel 153. The second vibration generator 151 may be provided on the rear side of the substrate 10. The second vibration generator 151 may be an IDT electrode. The IDT electrode may include a plurality of bars 1511 and 1512 facing each other and a plurality of fingers 1513 protruding from the bars 1511 and 1512. The sound wave may be a surface acoustic wave (SAW).

The separation unit 150 may separate a plurality of substances contained in the third liquid according to their molecular weight based on sound waves. The separator 150 may generate vibration around a channel through which liquid flows and separate a plurality of materials based on the vibration. For example, the separation unit 150 may separate a material bound to the target material (eg, a metal nanoparticle bound to an antibody) from a material not bound to the target material. The separator 150 may obtain the fourth liquid by filtering substances that are not combined with the target substance. The fourth liquid may be a liquid in which the concentration of substances not bound to the target substance in the third liquid is reduced.

The separator 150 may deliver the fourth liquid to the third chamber 33 through the first outlet channel 152. The separator 150 may deliver a substance that is not combined with the target substance to the fourth chamber 34 through the second outlet channel 153 .

The enrichment unit 160 may perform concentration processing on the fourth liquid contained in the third chamber 33. The concentrator 160 may allow the remaining materials to pass without passing the metal nanoparticle lumps containing the target material among the plurality of materials included in the fourth liquid. Accordingly, the concentrator 160 can obtain a fifth liquid with an improved concentration of metal nanoparticles or target materials.

Meanwhile, multiple biomarkers (or target substances) may be required to obtain diagnostic results for a subject. Hereinafter, a liquid purification device for purifying liquid containing a plurality of biomarkers will be described.

Figure 5 is a view from the top of a liquid purification device according to an embodiment of the present disclosure.

Referring to FIG. 5, the liquid purification device 500 includes a loading unit 110, an anticoagulation unit 120, a filter unit 130, a first reaction unit 540, a second reaction unit 542, and a first reaction unit 540. It may include a separation unit 551, a second separation unit 552, a first enrichment unit 561, and a second enrichment unit 562. Meanwhile, the loading unit 110, the anticoagulation unit 120, and the filter unit 130 have been described with reference to FIGS. 3 and 4, so duplicate descriptions will be omitted. In addition, the basic operations of the first reaction unit 541 and the second reaction unit 542 can be clearly understood through the reaction unit 140 of FIG. 3, and the first separation unit 551 and the second separation unit ( The basic operation of 552) can be clearly understood through the separation unit 150 of FIG. 3, and the basic operation of the first enrichment unit 561 and the second enrichment unit 562 can be understood through the enrichment unit 560 of FIG. 3. can be clearly understood through Therefore, descriptions overlapping with FIG. 3 are omitted.

The first reaction unit 541 may include a first reaction material storage unit 543 and a first reaction channel 544. The first reactive material storage unit 543 may store a first reactive material for detecting the first target material. The first reaction channel 544 may increase the mixing ratio of the first target material and the first reaction material. The first reaction channel 544 may induce an antigen-antibody reaction between the first target substance and the first reaction substance.

The second reaction unit 542 may include a second reaction material storage unit 543 and a first reaction channel 544. The second reactive material storage unit 545 may store a second reactive material for detecting the second target material. The second reaction channel 546 may increase the mixing ratio of the second target material and the second reaction material. The second reaction channel 546 may induce an antigen-antibody reaction between the second target substance and the second reaction substance.

The first liquid that has passed through the first reaction unit 541 contains a first metal nanoparticle mass formed by a reaction between the first target material and the first reaction material and a plurality of first materials that are not combined with the first target material. It can be included. The second liquid that has passed through the second reaction unit 542 contains a second metal nanoparticle mass formed by a reaction between the second target material and the second reaction material and a plurality of second materials that are not combined with the second target material. It can be included.

The first separation unit 551 may separate the first metal nanoparticle lump and the plurality of first materials contained in the first liquid. The first separator 551 may transfer the first metal nanoparticle mass to the third chamber 33 and the plurality of first materials to the fourth chamber 34. The first enrichment unit 561 may perform concentration treatment on the first metal nanoparticle mass accommodated in the third chamber 33.

The second separation unit 552 may separate the second metal nanoparticle lump contained in the second liquid and the plurality of second materials. The second separator 552 may deliver a second metal nanoparticle mass to the sixth chamber 36 and a plurality of second materials to the seventh chamber 37. The second enrichment unit 562 may perform concentration treatment on the second metal nanoparticle mass accommodated in the sixth chamber 36. Through the concentration process in the second enrichment unit 562, at least one material can be separated from the second metal nanoparticles and accommodated in the eighth chamber 38.

Figure 6 is a diagram showing a filter unit according to an embodiment of the present disclosure.

Referring to FIG. 6 , the filter unit 130 may include a first filter channel 131 and a fluid structure 132 protruding to the outside of the first filter channel 131 . The fluid structure 132 may be formed to form an obtuse angle with the direction (x) of the first filter channel 131. That is, the angle A1 between the direction (x) of the first filter channel 131 and the protrusion direction (y) of the fluid structure 132 may be 90 degrees to 180 degrees.

The first liquid 60 flowing into the first filter channel 131 may include a first material 61 and a second material 62. The first substance 61 may be one of the blood cell components, and the second substance 62 may be one of the plasma components.

The fluid structure 132 may include an air pocket 1321 through which the first liquid 60 does not flow. The air bladder 1321 may function as a pump. The air bag 1321 may repeat contraction and expansion based on the first sound wave generated by the first vibration generator 133. Accordingly, the air bag 1321 can pump the first liquid 60 in the direction (x) of the first filter channel 131.

Fluid structure 132 may function as a filter. When the first sound wave having a frequency corresponding to the size of the first material 61 is generated from the first vibration generator 133, the periphery of the interface 1322 of the first filter channel 131 and the air bladder 1321 A vortex 63 of the first liquid 60 may be generated in the area. The first material 61 may be trapped in the vortex 63. Accordingly, the fluid structure 132 may filter the first material 61.

Figure 7 is a diagram showing a separation unit according to an embodiment of the present disclosure.

Referring to FIG. 7 , the separator 150 may include a channel 70 through which the liquid 71 flows and a second vibration generator 151 located on the side of the channel 70 to generate sound waves. The separation unit 150 may include a first outlet channel 152 and a second outlet channel 153 branched from the channel 70 . In FIG. 7, the separator 150 is shown as having a symmetrical structure, but it is not limited thereto and may have an asymmetrical structure. The separation unit 150 may include three or more outlet channels.

The liquid 71 may include a target material 72, a reactive material 73 that reacts with the target material 72, and metal nanoparticles 75 to which the reactive material 73 and Raman active molecules 74 are combined. there is. A metal nanoparticle mass 76 may be formed by an antigen-antibody reaction between the target material 72 and the reactive material 73. The target material 72 may react with a plurality of reactive materials 73. The metal nanoparticle lump 76 may include a plurality of metal nanoparticles.

The separation unit 150 may separate a plurality of substances contained in the liquid 71 according to their molecular weight based on sound waves. A plurality of substances contained in the liquid 71 may be aligned according to their molecular weight when they encounter sound waves. At this time, the plurality of materials may move to form an acute angle with the direction of the channel 70.

The separator 150 may transfer the first material whose molecular weight is greater than a preset value to the third chamber 33 through the first outlet channel 152. The separator 150 may transfer the second material whose molecular weight is smaller than a preset value to the fourth chamber 34 through the second outlet channel 153. The first material may include a metal nanoparticle mass 76 formed by an antigen-antibody reaction between the target material 72 and the reactive material 73. The second material may include various materials other than the target material 72 or a material to which the target material 72 is combined. For example, the second material may include a reactive material 73 that does not react with the target material 72 . Alternatively, the second material may include metal nanoparticles and Raman active molecules bound to the reactive material 73 that has not reacted with the target material 72 .

FIG. 8 is a diagram for explaining a method of acquiring liquid information by a liquid information acquisition device according to an embodiment of the present disclosure.

Referring to FIG. 8, a substrate 82 onto which the liquid 81 is applied may be prepared. The liquid 81 may be one of the liquids stored in the plurality of chambers 31, 32, and 33 of the liquid purification device 100. The light source 210 of the liquid information acquisition device 200 may irradiate light to the liquid 81 applied to the substrate 82. The sensor 220 can sense light 84 scattered from fine particles contained in the liquid 81. The liquid information acquisition device 200 may acquire spectral information about the liquid 81 based on the scattered light 84.

A metal pattern 83 may be formed on the substrate 82 for surface enhanced Raman scattering. The scattered light 84 sensed by the sensor 220 by the metal pattern 83 may be amplified.

Liquid information acquisition methods for a plurality of liquids collected from the plurality of chambers 31, 32, and 33 may be different. The first liquid stored in the first chamber 31 and the second liquid stored in the second chamber 32 may not contain metal nanoparticles for surface enhanced scattering. Accordingly, when liquid information about the first liquid and the second liquid is obtained, the first liquid and the second liquid may be applied to the substrate 82 on which the metal pattern 83 is formed. The third liquid stored in the third chamber 33 may include metal nanoparticles. The third liquid may be applied to the substrate 82 on which the metal pattern 83 is not formed. Alternatively, in order to further improve the surface enhanced scattering effect, the third liquid may be applied to the substrate 82 on which the metal pattern 83 is formed.

Figure 9 is a block diagram showing the configuration of a diagnostic device according to an embodiment of the present disclosure. Referring to FIG. 9 , the diagnostic device 300 may include a communication interface 310, a memory 320, and a processor 330.

The communication interface 310 includes at least one communication circuit and can communicate with various types of external devices or external servers. For example, the communication interface 310 may receive information about the spectrum corresponding to the blood of the subject from an external device.

The communication interface 310 is a Wi-Fi communication module, a cellular communication module, a 3G (3rd generation) mobile communication module, a 4G (4th generation) mobile communication module, a 4th generation LTE (Long Term Evolution) communication module, and a 5G (5th generation) mobile communication module. It may include at least one of a module and wired Ethernet.

The memory 320 may store an operating system (OS) for controlling the overall operation of the components of the diagnostic device 300 and commands or data related to the components of the diagnostic device 300. The memory 320 may store information about a plurality of neural network models. Information about the plurality of neural network models may include information about parameters corresponding to each of the plurality of neural network models and learning data for learning the plurality of neural network models. The learning data may include label data (or ground truth) indicating the concentration corresponding to the intensity of the spectrum. The learning data may include label data indicating a diagnostic result corresponding to the intensity or concentration of the spectrum. The memory 320 may be implemented as non-volatile memory (ex: hard disk, solid state drive (SSD), flash memory) or volatile memory.

The processor 330 is electrically connected to the memory 320 and can control the overall functions and operations of the diagnostic device 300. The processor 330 may receive spectrum information (or information about the spectrum) about the liquid collected from the subject from an external device through the communication interface 310. Alternatively, the processor 330 may obtain spectrum information through an input interface. Information about the spectrum may include a number representing the spectrum, a vector, and at least one peak value included in the spectrum. The frequency at which the peak value of the spectrum appears may correspond to the Raman active molecule bound to the metal nanoparticle bound to the target material.

The processor 330 may obtain information about the spectrum of a plurality of liquids. For example, the processor 330 may obtain information about the spectrum corresponding to the first liquid collected from the subject. The first liquid may be whole blood of the subject. The processor 330 may obtain information about the spectrum corresponding to the second liquid in which the concentration of blood cell components has been reduced in the first liquid. The processor 330 acquires information about the spectrum corresponding to the third liquid to which the first reactive material (e.g., first antibody) for detection of the first target material (e.g., first antigen) is added to the second liquid. can do. The processor 330 acquires information about the spectrum corresponding to the fourth liquid to which a second reaction material (e.g., a second antibody) for detection of a second target material (e.g., a second antigen) is added to the second liquid. can do.

The processor 330 may obtain concentration information of the target material based on information about the spectrum. The concentration information of the target material may include a numerical value representing the concentration of the target material, a feature vector corresponding to the concentration of the target material, and a peak value of a spectrum corresponding to the concentration of the target material.

The processor 330 may obtain first concentration information of the first target material based on information about the first spectrum and information about the second spectrum. Here, the information about the first spectrum may be information about the first liquid that includes the first target material and does not include the first reactive material that reacts with the first target material. The information about the second spectrum may be information about the second liquid in which the first reaction material is added to the first liquid. For example, the first liquid may be the liquid contained in the second chamber 32 of FIG. 3 . The second liquid may be the liquid contained in the third chamber 33 of FIG. 3. The second liquid may have a lower content of components other than the first target material (eg, proteins other than the first target material) compared to the first liquid. The remaining components may act as noise when obtaining information about the spectrum. Accordingly, the processor 330 uses information on the spectrum of each of the first liquid from which the remaining components have not been removed and the second liquid from which at least a portion of the remaining components have been removed, thereby providing first concentration information from which noise due to the remaining components has been reduced. can be obtained.

The processor 330 may obtain concentration information based on a lookup table in which information about the second spectrum and the peak value and concentration information of the spectrum are matched. The lookup table may be previously stored in the memory 320. For example, the processor 330 may identify concentration information corresponding to the peak value of the second spectrum in the lookup table.

The processor 330 may correct the obtained density information based on the lookup table. For example, the processor 330 may input information about the first spectrum into the first neural network model to obtain a coefficient for correcting the concentration information. The processor 330 may obtain first concentration information of the first target material by correcting the concentration information based on the coefficient.

The processor 330 may obtain first concentration information about the first target material by inputting information about the first spectrum and information about the second spectrum into the second neural network model. The second neural network model may be a model learned to obtain concentration information based on information about the spectrum.

The processor 330 may obtain a feature vector by inputting information about the second spectrum into the second neural network model. The processor 330 may obtain first concentration information by inputting information about the feature vector and the first spectrum into a third neural network model. The third neural network model may be a model learned to obtain corrected concentration information based on information about the spectrum and concentration information. The feature vector can be obtained at the output of a layer (eg, fully connected layer) included in the second neural network model.

The processor 330 may obtain a first feature vector by inputting information about the first spectrum into a second neural network model. The processor 330 may obtain a second feature vector by inputting information about the second spectrum into a second neural network model. The processor 330 may obtain first concentration information by inputting the first feature vector and the second feature vector into the fourth neural network model. The fourth neural network model may be a model learned to obtain corrected concentration information based on the concentration information.

The processor 330 may obtain diagnostic information based on spectrum information and concentration information. For example, the processor 330 may obtain diagnostic information by inputting information about the first spectrum and first concentration information into the fifth neural network model. Alternatively, the processor 330 may obtain diagnostic information by inputting information about the first spectrum, first concentration information of the first target material, and second concentration information of the second target material into the fifth neural network model. The fifth neural network model may be a model learned to obtain diagnostic information based on spectrum information and concentration information.

The processor 330 may obtain diagnostic information based on information about spectra corresponding to a plurality of liquids. For example, the processor 330 may obtain diagnostic information by inputting information about the first spectrum and information about the second spectrum into a sixth neural network model. The sixth neural network model may be a model learned to obtain diagnostic information based on spectrum information.

Meanwhile, functions related to artificial intelligence according to the present disclosure are operated through the processor 330 and memory 320. The processor 330 may be comprised of one or multiple processors. At this time, one or more processors may be a general-purpose processor such as a CPU, AP, or DSP (Digital Signal Processor), a graphics-specific processor such as a GPU or VPU (Vision Processing Unit), or an artificial intelligence-specific processor such as an NPU. One or more processors control input data to be processed according to predefined operation rules or artificial intelligence models stored in the memory 320. Alternatively, when one or more processors are dedicated artificial intelligence processors, the artificial intelligence dedicated processors may be designed with a hardware structure specialized for processing a specific artificial intelligence model.

Predefined operation rules or artificial intelligence models are characterized by being created through learning. Here, being created through learning means that the basic artificial intelligence model is learned using a large number of learning data by a learning algorithm, thereby creating a predefined operation rule or artificial intelligence model set to perform the desired characteristics (or purpose). It means burden. This learning may be accomplished in the device itself that performs the artificial intelligence according to the present disclosure, or may be accomplished through a separate server and/or system. Examples of learning algorithms include supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but are not limited to the examples described above.

Artificial intelligence models can be created through learning. An artificial intelligence model may be composed of multiple neural network layers. Each of the plurality of neural network layers has a plurality of weight values, and neural network calculation is performed through calculation between the calculation result of the previous layer and the plurality of weights. Multiple weights of multiple neural network layers can be optimized by the learning results of the artificial intelligence model. For example, a plurality of weights may be updated so that loss or cost values obtained from the artificial intelligence model are reduced or minimized during the learning process.

Artificial neural networks may include deep neural networks (DNN), such as Convolutional Neural Network (CNN), Deep Neural Network (DNN), Recurrent Neural Network (RNN), Generative Adversarial Network (GAN), Examples include Restricted Boltzmann Machine (RBM), Deep Belief Network (DBN), Bidirectional Recurrent Deep Neural Network (BRDNN), or Deep Q-Networks, but are not limited to the examples described above.

Figure 10 is a diagram showing information on a spectrum according to an embodiment of the present disclosure.

Referring to FIG. 10, information about the spectrum (S) (or spectrum information) may mean the intensity of a Raman signal according to wavenumber. Spectrum information (S) may be vector-type data in which the wavenumber and the intensity of the Raman signal are matched to each other. In various embodiments described later, spectrum information (S) may be input to a neural network model. At this time, the spectrum information (S) input to the neural network model may include a plurality of intensities each corresponding to a plurality of wave numbers. Alternatively, the spectrum information (S) may mean a single intensity corresponding to a specific wave number (eg, a wave number corresponding to a Raman active molecule).

Spectral information (S) may be normalized data. For example, a first Raman signal may be obtained by irradiating a laser beam of first intensity to a specific liquid, and a second Raman signal may be obtained by irradiating a laser beam of second intensity. At this time, the diagnostic device 300 may adjust the intensity of the first Raman signal or the second Raman signal so that the area of the first Raman signal and the area of the second Raman signal are the same.

Figure 11 is a diagram for explaining a method of obtaining concentration information according to the first embodiment of the present disclosure.

Referring to FIG. 11 , the diagnostic device 300 may acquire information (S1) about the first spectrum corresponding to the first liquid and information (S2) about the second spectrum corresponding to the second liquid. The first liquid may include a first target material (eg, antigen A) and may not contain a first reactive material (eg, antibody A') corresponding to the first target material. The second liquid is a liquid in which a first reactive material is added to the first liquid, and may include a first target material combined with the first reactive material.

The diagnostic device 300 may acquire first concentration information 303 of the first target material based on information about the second spectrum (S2) and the look-up table 302. The look-up table 302 may include peak values and concentrations of spectra that match each other. The diagnostic device 300 may obtain the peak value of the second spectrum based on the information (S2) about the second spectrum. For example, the diagnostic device 300 may identify the intensity corresponding to the preset wavenumber in the information about the second spectrum (S2) as the peak value. Here, the preset wave number may correspond to the Raman active molecule bound to the metal nanoparticle bound to the first reactant.

The diagnostic device 300 may obtain the first concentration information 303 based on the peak value obtained from the information on the second spectrum (S2) and the look-up table 302. For example, the diagnostic device 300 may obtain the first concentration information 303 by identifying the concentration corresponding to the peak value obtained in the information about the second spectrum (S2) in the look-up table 302. there is.

Meanwhile, the second liquid may contain various other components (eg, proteins, etc.) in addition to the first target material. Accordingly, the first concentration information 303 obtained based on the look-up table 302 may include an error.

The diagnostic device 300 may correct the first concentration information 303 based on the information about the first spectrum (S1). The diagnostic device 300 may obtain a coefficient 301 for correcting concentration information by inputting information about the first spectrum (S1) into the first neural network model (M1). At this time, the information S1 about the first spectrum may include a plurality of intensities according to a plurality of wavenumbers. The diagnostic device 300 may correct the first concentration information 303 based on the coefficient 301. For example, the diagnostic device 300 may obtain the second concentration information 304 by multiplying the first concentration information 303 by the coefficient 301.

FIG. 12 is a diagram for explaining a method of learning a first neural network model according to an embodiment of the present disclosure.

Referring to FIG. 12, the training data of the first neural network model (M1) may include a plurality of first spectrum information (S1), a plurality of second spectrum information (S2), and a first ground truth (GT1). . The first ground truth GT1 may include density information of the target material corresponding to spectrum information corresponding to the liquid containing the target material. The first ground truth GT1 may be previously stored in the memory 320 of the diagnostic device 300.

The diagnostic device 300 may learn the first neural network model (M1) based on the first ground truth (GT1). The diagnostic device 300 may acquire second concentration information 304 based on the first spectrum information S1 and the second spectrum information S2. The method of acquiring the second concentration information 304 is described in FIG. 11 and detailed description will be omitted. The diagnostic device 300 may obtain the first loss value based on the second concentration information 304 and the concentration information included in the first ground truth GT1. The diagnostic device 300 may update parameters (eg, weights) of the first neural network model (M1) until the first loss value becomes smaller than a preset value. That is, the diagnostic device 300 can learn the first neural network model (M1) based on backpropagation.

Meanwhile, in FIG. 12, it is shown as an example that the diagnostic device 300 learns the first neural network model (M1) based on supervised learning, but this is only an example, and the diagnostic device 300 uses unsupervised learning (for example, , reinforcement learning), the first neural network model (M1) may be trained.

Figure 13 is a diagram for explaining a method of obtaining concentration information according to a second embodiment of the present disclosure.

Referring to FIG. 13, the diagnostic device 300 inputs information about the first spectrum (S1) and information about the second spectrum (S2) into the second neural network model (M2) to obtain concentration information of the first target material ( 311) can be obtained. For example, information about the second spectrum (S2) may mean a single intensity corresponding to a specific wave number (eg, a wave number corresponding to a Raman active molecule).

The first liquid and the second liquid may commonly include other components (eg, blood cell components) in addition to the first target material. The second neural network model (M2) can acquire concentration information 311 based on spectral information corresponding to a plurality of liquids that commonly contain other components, rather than based only on spectral information corresponding to one liquid. . Accordingly, the accuracy of the concentration information 311 can be improved.

The diagnostic device 300 may learn the second neural network model (M2). The training data of the second neural network model (M2) may include a plurality of first spectrum information (S1), a plurality of second spectrum information (S2), and a second ground truth. The second ground truth may include concentration information corresponding to spectrum information. The diagnostic device 300 may obtain the second loss value based on the concentration information 311 and the concentration information included in the second ground truth. The diagnostic device 300 may update parameters (eg, weights) of the second neural network model (M2) until the second loss value becomes smaller than a preset value.

Figure 14 is a diagram for explaining a method of obtaining concentration information according to a third embodiment of the present disclosure.

Referring to FIG. 14, the diagnostic device 300 may obtain a feature vector 321 by inputting information about the second spectrum (S2) into a second neural network model (M2). The feature vector 321 may be output through the fully connected layer (FC) of the second neural network model (M2). The diagnostic device 300 may obtain concentration information 322 by inputting information about the first spectrum (S1) and the feature vector 321 into the third neural network model (M3).

The diagnostic device 300 can learn the third neural network model (M3). The training data of the third neural network model (M3) may include a plurality of first spectrum information (S1), a plurality of second spectrum information (S2), and a third ground truth. The third ground truth may include spectrum information and concentration information corresponding to the feature vector. The diagnostic device 300 may obtain the third loss value based on the concentration information 322 and the concentration information included in the third ground truth. The diagnostic device 300 may update parameters (eg, weights) of the third neural network model (M3) until the third loss value becomes smaller than a preset value.

The diagnostic device 300 may also train the second neural network model (M2) while training the third neural network model (M3). The diagnostic device 300 may update the parameters of the second neural network model (M2) until the third loss value becomes smaller than a preset value. Alternatively, the diagnostic device 300 may train the third neural network model (M3) after training of the second neural network model (M2) is completed. At this time, while the parameters of the third neural network model (M3) are updated, the parameters of the second neural network model (M2) may be frozen.

Figure 15 is a diagram for explaining a method of obtaining concentration information according to a fourth embodiment of the present disclosure.

Referring to FIG. 15 , the diagnostic device 300 may obtain a first feature vector 331 by inputting information about the first spectrum (S1) into a second neural network model (M2). The diagnostic device 300 may acquire a second feature vector 332 by inputting information about the second spectrum (S2) into a second neural network model (M2). The diagnostic device 300 may obtain concentration information 333 by inputting the first feature vector 331 and the second feature vector 332 into the fourth neural network model (M4).

The diagnostic device 300 can learn the fourth neural network model (M4). The training data of the fourth neural network model (M4) may include a plurality of first spectrum information (S1), a plurality of second spectrum information (S2), and a fourth ground truth. The fourth ground truth may include density information corresponding to the feature vector. The diagnostic device 300 may obtain the fourth loss value based on the concentration information 333 and the concentration information included in the fourth ground truth. The diagnostic device 300 may update parameters (eg, weights) of the fourth neural network model (M4) until the fourth loss value becomes smaller than a preset value.

The diagnostic device 300 may also train the second neural network model (M2) while training the fourth neural network model (M4). The diagnostic device 300 may update the parameters of the second neural network model (M2) until the fourth loss value becomes smaller than a preset value. Alternatively, the diagnostic device 300 may train the fourth neural network model (M4) after training of the second neural network model (M2) is completed. At this time, while the parameters of the fourth neural network model (M4) are updated, the parameters of the second neural network model (M2) may be frozen.

Figure 16 is a diagram for explaining a method of obtaining concentration information according to an embodiment of the present disclosure.

Referring to FIG. 16, the diagnostic device 300 obtains diagnostic information 342 by inputting information about the first spectrum (S1) and concentration information 341 of the first target material into the fifth neural network model (M5). can do. For example, the diagnostic information 342 may include information about a disease that the subject is expected to have. At this time, the first target material may be a biomarker of the disease.

The diagnostic device 300 can learn the fifth neural network model (M5). The training data of the fifth neural network model (M5) may include a plurality of first spectrum information (S1), a plurality of concentration information (341), and a fifth ground truth. A plurality of concentration information 341 can be obtained according to the various embodiments described above. The fifth ground truth may include diagnostic information corresponding to spectrum information and concentration information.

The diagnostic device 300 may obtain the fifth loss value based on the diagnostic information included in the diagnostic information 342 and the fifth ground truth. The diagnostic device 300 may update parameters (eg, weights) of the fifth neural network model (M5) until the fifth loss value becomes smaller than a preset value. The diagnostic device 300 may train the neural network model that outputs the concentration information 341 while training the fifth neural network model (M5). For example, when the concentration information 341 is acquired by the second neural network model (M2), the diagnostic device 300 uses the second neural network model (M2) until the fifth loss value becomes smaller than the preset value. Parameters can be updated. Alternatively, the diagnostic device 300 may train the fifth neural network model (M5) after training of the neural network model that outputs the concentration information 341 is completed.

Meanwhile, disease diagnosis may be performed based on one biomarker, but may also be performed using multiple biomarkers.

Figure 17 is a diagram for explaining a method of obtaining concentration information according to an embodiment of the present disclosure.

Referring to FIG. 17, the diagnostic device 300 combines information about the first spectrum (S1), first concentration information 351 of the first target material, and second concentration information 352 of the second target material into the fifth Diagnostic information 353 can be obtained by inputting it into the neural network model (M5). For example, the diagnosis information 353 may include a diagnosis result for Alzheimer's disease. The first target material may be tau protein, and the second target material may be beta-amyloid.

Figure 18 is a diagram for explaining a method of obtaining concentration information according to an embodiment of the present disclosure.

Referring to FIG. 18, the diagnostic device 300 obtains diagnostic information 361 by inputting information about the first spectrum (S1) and information about the second spectrum (S2) into the sixth neural network model (M6). You can. Information (S1) about the first spectrum may correspond to the first liquid containing the first target material. The information S2 about the second spectrum may correspond to the second liquid including the first target material and the first reactive material combined with the first target material.

The diagnostic device 300 can learn the sixth neural network model (M6). The training data of the sixth neural network model (M6) may include a plurality of first spectrum information (S1), a plurality of second spectrum information (S2), and a sixth ground truth. The sixth ground truth may include diagnostic information corresponding to spectrum information. The diagnostic device 300 may obtain the sixth loss value based on the diagnostic information included in the diagnostic information 361 and the sixth ground truth. The diagnostic device 300 may update parameters (eg, weights) of the sixth neural network model (M6) until the sixth loss value becomes smaller than a preset value.

Preprocessing may be performed on input data input to a plurality of neural network models (M1, M2, M3, M4, M5, M6) according to the present disclosure. For example, concatenation may be performed on a plurality of input data.

Some of the multiple neural network models (M1, M2, M3, M4, M5, M6) may be integrated.

Figure 19 is a flowchart showing a control method of a diagnostic device according to an embodiment of the present disclosure.

Referring to FIG. 19, the diagnostic device 300 may obtain information about the first spectrum corresponding to the first liquid collected from the subject and information about the second spectrum corresponding to the second liquid (S1910). . The first liquid may include a first target material. The second liquid may be a liquid in which a first reactive material corresponding to the first target material is added to the first liquid. The second liquid may include a first target material combined with a first reactant.

The diagnostic device 300 may acquire first concentration information of the first target material based on information about the first spectrum and information about the second spectrum (S1920). For example, the diagnostic device 100 may obtain concentration information based on a look-up table in which information about the second spectrum and peak values of the spectrum and concentration information are matched. The diagnostic device 100 may obtain a coefficient for correcting concentration information by inputting information about the first spectrum into the first neural network model. The diagnostic device 300 may obtain corrected concentration information by performing correction on the concentration information based on the coefficient.

The diagnostic device 300 may obtain first concentration information by inputting information about the first spectrum and information about the second spectrum into a second neural network model.

The diagnostic device 300 may obtain a feature vector by inputting information about the second spectrum into a second neural network model. The diagnostic device 300 may obtain first concentration information by inputting information about the first spectrum and feature vectors into a third neural network model.

The diagnostic device 300 may obtain a first feature vector by inputting information about the first spectrum into a second neural network model. The diagnostic device 300 may obtain a second feature vector by inputting information about the second spectrum into a second neural network model. The diagnostic device 300 may obtain first concentration information by inputting the first feature vector and the second feature vector into the fourth neural network model.

The diagnostic device 300 may obtain diagnostic information about the subject based on the first concentration information (S1930). The diagnostic device 300 may obtain diagnostic information by inputting information about the first spectrum and first concentration information into the fifth neural network model.

The diagnostic device 300 may obtain information about a third spectrum corresponding to a third liquid in which a second reactive material is added to the first liquid. The third liquid may include a second target material combined with a second reactant. The diagnostic device 300 may obtain second concentration information of the second target material based on information about the first spectrum and information about the third spectrum. The diagnostic device 300 may obtain diagnostic information by inputting information about the first spectrum, first concentration information, and second concentration information into the fifth neural network model.

The diagnostic device 300 may obtain diagnostic information by inputting information about the first spectrum and information about the second spectrum into a sixth neural network model.

The various embodiments described above may be implemented in a recording medium that can be read by a computer or similar device using software, hardware, or a combination thereof. In some cases, embodiments described herein may be implemented with a processor itself. When implemented as software, embodiments such as procedures and functions described in this specification may be implemented as separate software modules. Each of the software modules may perform one or more functions and operations described herein.

Computer instructions for performing processing operations according to various embodiments of the present disclosure described above may be stored in a non-transitory computer-readable medium. Computer instructions stored in such a non-transitory computer-readable medium, when executed by a processor, can cause a specific device to perform processing operations according to the various embodiments described above.

A non-transitory computer-readable medium refers to a medium that stores data semi-permanently and can be read by a device, rather than a medium that stores data for a short period of time, such as registers, caches, and memories. Specific examples of non-transitory computer-readable media may include CD, DVD, hard disk, Blu-ray disk, USB, memory card, ROM, etc.

A storage medium that can be read by a device may be provided in the form of a non-transitory storage medium. Here, 'non-transitory storage medium' simply means that it is a tangible device and does not contain signals (e.g. electromagnetic waves). This term refers to cases where data is semi-permanently stored in a storage medium and temporary storage media. It does not distinguish between cases where it is stored as . For example, a 'non-transitory storage medium' may include a buffer where data is temporarily stored.

Methods according to various embodiments disclosed in this document may be provided and included in a computer program product. Computer program products are commodities and can be traded between sellers and buyers. The computer program product may be distributed in the form of a machine-readable storage medium (e.g. compact disc read only memory (CD-ROM)) or through an application store (e.g. Play StoreTM) or on two user devices (e.g. It can be distributed (e.g. downloaded or uploaded) directly between smartphones) or online. In the case of online distribution, at least a portion of the computer program product (e.g., a downloadable app) is stored on a machine-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server. It can be temporarily stored or created temporarily.

In the above, preferred embodiments of the present disclosure have been shown and described, but the present disclosure is not limited to the specific embodiments described above, and may be used in the technical field pertaining to the disclosure without departing from the gist of the disclosure as claimed in the claims. Of course, various modifications can be made by those with the knowledge, and these modifications should not be understood individually from the technical ideas or perspectives of the present disclosure.

100: liquid purification device 200: liquid information acquisition device
300: diagnostic device 1000: diagnostic system

Claims (1)

In the liquid purification device,
Board;
a loading portion formed on an upper portion of the substrate and containing a first liquid;
a filter unit that reduces the concentration of at least one component contained in the first liquid to obtain a second liquid in which the concentration of the at least one substance is reduced;
The second liquid is mixed with a reactive material for detecting the target material, so that among the plurality of materials included in the second liquid, a first material that undergoes a predetermined reaction with the reactive material and a first material that does not undergo a predetermined reaction with the reactive material a reaction unit that obtains a third liquid containing a second substance that is not present; and
A separation unit for separating the first material and the second material; comprising
Liquid purification device.

Images