KR101927852B1 - Method and Apparatus for Identifying Cell Species Using 3D Refractive Index Tomography and Machine Learning Algorithm - Google Patents

Abstract

A method and apparatus for classifying non-stained non-labeled cells using 3D refractive tomography and dip learning are presented. The method of distinguishing non-stained non-labeled cells using a three-dimensional refractive tomogram and a deep-running method comprises: optically measuring a three-dimensional refractive index distribution of a cell; And classifying the type of the cell with an un-stained non-label by applying a deep learning algorithm or a convolutional neural network algorithm using the measured three-dimensional refractive index distribution .

Description

TECHNICAL FIELD The present invention relates to a method and apparatus for distinguishing non-stained non-labeled cells using three-dimensional refractive index tomograms and deep-

The following examples relate to a method and apparatus for distinguishing non-stained non-labeled cells using three-dimensional refractive tomograms and dip-runs, and more particularly, The present invention relates to a method and an apparatus for classifying non-stained non-labeled cells using a three-dimensional refractive tomogram and a deep-run.

Cell sorting is a very important issue in life sciences and medicine. For example, the division of B cells and T cells, which are subtypes of white blood cells, is an essential process for immunotherapy.

Currently, cell staining or cell-specific antigen-antibody reactions are generally used to distinguish the types of cells. The cell staining method uses a fluorescent protein, a dye, a quantum dot, etc. to visualize and visualize specific organelles in the cell, and to identify cells based on the user's experience. Therefore, it takes time and expense in the dyeing process, and there is a fundamental problem that the cells are not able to be injected into the body again due to the inevitable deformation of the cells due to the dyeing process. There is also a problem of poor accuracy because cells are identified based on user experience.

The method using the cell-specific antigen-antibody reaction is a method of using an antibody capable of binding to a protein specifically present in a cell membrane or the like. For the purpose of measurement and discrimination, it becomes possible to recognize a specific cell using a fluorescent substance or a magnetic substance in the antibody.

However, the process of producing antibodies is very expensive, and biocompatibility issues have been raised in immunotherapy or stem cell applications where additional fluorescent / magnetic substances attach to the cells from the outside and must be injected back into the body.

Korean Patent Laid-Open Publication No. 10-2012-0087165 relates to the detection, measurement and imaging of cells such as cancer and other biological materials using such targeted nanoparticles and their magnetic properties, in which targeted magnetic nanoparticles and special magnetic systems Describes techniques and apparatuses for detecting, measuring, or locating cells or materials that are present in a subject even at a very low concentration in vivo.

1. Kim, K., et al. (2016). "Optical diffraction tomography techniques for the study of cell pathophysiology." arXiv preprint arXiv: 1603.00592. 2. Lee, K., et al. (2013). "Quantitative phase imaging techniques for the study of cell pathophysiology: from principles to applications." Sensors 13 (4): 4170-4191. 3. Wolf, E. (1969). "Three-dimensional structure determination of semi-transparent objects from holographic data." Optics Communications 1 (4): 153-156. 4. Krizhevsky, A., et al. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems.

Embodiments describe methods and apparatus for differentiating non-stained non-stained cells using three-dimensional refractive tomograms and dip-runs, and more specifically, using three-dimensional refractive index measurements of living cells without using staining or labeling, Provides a technique for classifying.

The embodiments optically measure the three-dimensional refractive index distribution of cells and apply deep learning or convolutional neural network algorithms to classify the types of living cells with non-staining non-labeling The present invention also provides a method and apparatus for classifying non-stained non-labeled cells using a three-dimensional refractive index tomogram and a deep running.

A method for distinguishing un-stained unlabeled cell types using three-dimensional refractive index tomograms and dip-runs according to an embodiment includes: optically measuring a three-dimensional refractive index distribution of cells; And applying a deep learning algorithm or a convolutional neural network algorithm using the measured three-dimensional refractive index distribution to classify the type of the cell with an un-stained non-labeling.

The step of optically measuring the three-dimensional refractive index distribution of the cell may include the steps of: injecting light from the light source into the cell; Measuring a transmitted light diffracted by the cell with an interferometer to obtain a plurality of two-dimensional holograms; And measuring the three-dimensional refractive index profile of the cell using the plurality of two-dimensional holograms.

The step of optically measuring the three-dimensional refractive index profile of the cell comprises rotating the angle of light incident on the cell and measuring the three-dimensional refractive index profile of the cell using the plurality of two-dimensional holograms measured by the interferometer .

The step of optically measuring the three-dimensional refractive index profile of the cell may be performed by directly rotating the cell and measuring the three-dimensional refractive index profile of the cell using the plurality of two-dimensional holograms measured by the interferometer.

The step of optically measuring the three-dimensional refractive index profile of the cell can measure the three-dimensional refractive index profile of the cell through optical measurement of at least one of optical diffraction tomography and optical projection tomography.

The step of classifying the cell types may include extracting a refractive index characteristic specific to a cell type through the convolutional neural network algorithm using the measured three-dimensional refractive index distribution; And a classification step of classifying the cell type and species according to the refractive index characteristic specific to the extracted cell type.

The step of classifying the cell type may include extracting a refractive index characteristic specific to a cell type through the depth learning algorithm using the measured three-dimensional refractive index distribution; And a classification step of classifying the cell type and species according to the refractive index characteristic specific to the extracted cell type.

A non-stained non-labeled cell sorting apparatus using a three-dimensional refractive index tomogram and a dip running according to another embodiment includes a three-dimensional refractive index measuring unit for optically measuring a three-dimensional refractive index distribution of a cell; And a cell type determination unit for classifying the type of the cell by non-staining non-labeling by applying a deep learning or a convolutional neural network algorithm using the measured three-dimensional refractive index distribution.

The three-dimensional refractive index measuring unit may include: a light source that causes light to enter the cell; An interferometer for measuring transmitted light diffracted in the cell to obtain a plurality of two-dimensional holograms; And a measuring unit for measuring a three-dimensional refractive index distribution of the cell using the plurality of two-dimensional holograms, wherein the three-dimensional refractive index measuring unit measures the refractive index of the cell by using at least one of optical diffraction tomography and optical projection tomography The three-dimensional refractive index distribution of the cell can be measured.

The cell type determination unit may include a characteristic extraction unit that extracts a refractive index characteristic specific to a cell type through the deep learning algorithm or the convolutional neural network algorithm using the measured three- ; And a classification unit for classifying the cell type and species according to the refractive index characteristic specific to the extracted cell type.

According to the embodiments, the refractive index distribution of the three-dimensional cell is optically measured, and deep learning or convolutional neural network algorithm is applied to determine the type of living cells And a method and apparatus for distinguishing non-stained non-labeled cell types using deep running.

 According to the embodiments, classification of cells without using dyeing or labeling can be widely used in the field of cell biology, particularly in the field of cytometry, A method and apparatus for classifying non-stained non-labeled cells using a refractive index tomogram and a deep running can be provided.

FIG. 1 is a view for explaining a method of distinguishing non-stained non-labeled cells using a three-dimensional refractive index tomogram and a deep running according to an embodiment.
FIG. 2 is a view for explaining an un-stained unlabeled cell sorting apparatus using a three-dimensional refractive index tomogram and a dip running according to an embodiment.
3A is a view for explaining a method of measuring a three-dimensional refractive index of a cell using an incident light rotation method according to an embodiment.
FIG. 3B is a view for explaining a method of measuring the three-dimensional refractive index of cells using the cell rotation method according to one embodiment.
FIG. 4 is a diagram for explaining deep learning using a data set according to an embodiment and cell type classification. Referring to FIG.
FIGS. 5 to 7 are flowcharts for explaining an unoccluded unlabeled cell sorting method using a three-dimensional refractive index tomogram and dip learning according to an embodiment.

Hereinafter, embodiments will be described with reference to the accompanying drawings. However, the embodiments described may be modified in various other forms, and the scope of the present invention is not limited by the embodiments described below. In addition, various embodiments are provided to more fully describe the present invention to those skilled in the art. The shape and size of elements in the drawings may be exaggerated for clarity.

The following examples provide a way to distinguish the types of living cells without additional steps such as staining or labeling. More specifically, the embodiments optically measure the three-dimensional refractive index distribution of a cell, apply deep learning or a convolutional neural network algorithm, Can be distinguished.

For example, when a laser beam is irradiated from a light source to a cell (sample), a plurality of 2D images can be obtained while the laser beam is turned on, and a 3D image can be obtained by obtaining a 3D absorbance by performing a tomogram through analysis. Thus, three-dimensional images can be obtained without staining the cells. In the case of confirming 3-dimensional images of cells, it is necessary to stain the cells, but it is difficult to inject into the body after staining of the cells, which makes it difficult to apply to various fields.

FIG. 1 is a view for explaining a method of distinguishing non-stained non-labeled cells using a three-dimensional refractive index tomogram and a deep running according to an embodiment.

The present embodiments provide a method of classifying a cell type using a three-dimensional refractive index measurement of living cells without using dyeing or marking.

The intracellular 3D refractive index distribution is closely related to the composition and morphology of subcellular organelles. In addition, the refractive index value itself is proportional to the concentration of protein, which is a major component in the cell. Thus, measuring the three-dimensional refractive index information of a cell reflects not only morphological characteristics of cells and intracellular organelles but also biochemical characteristics information.

Accordingly, when the three-dimensional refractive index distribution is used, it is possible to classify the types of cells using specific morphological characteristics and biochemical characteristics of the cell types.

By measuring the three-dimensional refractive index of a cell and applying the measured value to a machine learning algorithm, the kind of living cell can be easily and accurately distinguished.

Referring to FIG. 1, a non-stained non-labeled cell type classification method using a three-dimensional refractive tomogram and a deep learning according to an embodiment measures a three-dimensional refractive index image of cells (110) 120). At this time, a large number of specimens (> 100) are measured for each type, and a refractive index characteristic (112) specific to the cell type can be extracted (112) by referring to the pre-stored vertical library 111 using a deep learning algorithm have. The type of the cell can then be identified 120 by classifying the cell type and species according to the extracted characteristics.

Hereinafter, a method and an apparatus for distinguishing un-stained unlabeled cell types using a three-dimensional refractive index tomogram and a dip running according to one embodiment will be described in more detail.

FIG. 2 is a view for explaining an un-stained unlabeled cell sorting apparatus using a three-dimensional refractive index tomogram and a dip running according to an embodiment.

Referring to FIG. 2, an apparatus for classifying non-stained unlabeled cell using a three-dimensional refractive index tomogram and a deep running according to an embodiment includes a three-dimensional refractive index measurement unit 210 and a cell type determination unit 220 .

Here, the three-dimensional refractive index measuring unit 210 may include a light source, an interferometer, and a measuring unit. The cell type determination unit 220 may include a characteristic extraction unit and a classification unit.

The three-dimensional refractive index measuring unit 210 can optically measure the three-dimensional refractive index distribution of the cell.

For example, the three-dimensional refractive index measurement unit 210 may be an optical system including a light source and a camera, and may have various shapes such as a reflection type and a transmission type.

The three-dimensional refractive index measurement unit 210 may include a light source, an interferometer, and a measurement unit.

The light source can cause light to enter the cell. That is, the light source can irradiate cells with light. For example, a laser can be used as a light source, and a light source can irradiate a laser beam onto a sample such as a cell to be measured.

Here, the cell may be represented by a sample representing an object to be measured, and it may be a bacterium or a microorganism as well as a cell, and may be an object including a cell or the like.

The light source used may utilize a single wavelength laser. In addition, the light source can utilize a larger amount of information to differentiate cells by measuring the refractive index of three dimensions at each wavelength using a multi-wavelength laser.

The interferometer can acquire a plurality of two-dimensional holograms by measuring the transmitted light diffracted by the cell after the light incident from the light source is incident on the cell.

Here, the interferometer is a measuring device using the interference phenomenon of light, which divides the light emitted from the same light source into two or more beams to make a difference in the propagation path, and then observes the interference phenomenon occurring when the light meets again.

The measuring unit can measure the three-dimensional refractive index distribution of the cell using a plurality of two-dimensional holograms obtained from the interferometer. For example, a camera, which is a photographing apparatus for photographing an image, may be used as the measuring unit.

The three-dimensional refractive index measuring unit 210 can measure the three-dimensional refractive index distribution of the cell through optical measurement of at least one of optical diffraction tomography and optical projection tomography.

The three-dimensional refractive index measuring unit 210 may measure the three-dimensional refractive index distribution of the cell using a plurality of two-dimensional holograms measured in the interferometer by rotating the angle of light incident on the cell.

For example, by using a wavefront controller to measure the angle of incidence on a cell while measuring several pieces of two-dimensional holograms, and analyzing the measured hologram information through an algorithm, a three-dimensional refractive index distribution of the cell can be measured. In particular, a high-speed three-dimensional optical tomography can be performed using a digital micromirror device, which is one of wavefront controllers.

The three-dimensional refractive index measuring unit 210 may also measure the three-dimensional refractive index distribution of the cells using a plurality of two-dimensional holograms measured in the interferometer by directly rotating the cells.

On the other hand, the method of measuring cells is a method in which cells are placed at a low concentration on an in vitro slide glass, a form in which cells are concentrated at a high concentration on an in vitro slide glass, Layered forms, tissue slides in which biopsy slides are cut to a thickness of between 5 micrometers and 50 micrometers, and microfluidic channels for high-throughput screening in vitro ). ≪ / RTI >

The cell type determination unit 220 applies deep learning or a convolutional neural network algorithm using the three-dimensional refractive index distribution measured by the three-dimensional refractive index measurement unit 210, The type of cells can be distinguished.

The cell type determination unit 220 may include a characteristic extraction unit and a classification unit.

The characteristic extracting unit can extract a refractive index characteristic specific to a cell type through a depth learning algorithm or a convolutional neural network algorithm using the measured three-dimensional refractive index distribution.

The classification unit can classify the cell type and species according to the refractive index characteristic specific to the cell type extracted from the characteristic extracting unit.

According to the embodiments, it is possible to provide a technique of sorting cells by using a three-dimensional refractive index measurement of living cells without using dyeing or labeling.

According to the embodiments, the refractive index distribution of the three-dimensional cell is optically measured, and deep learning or convolutional neural network algorithm is applied to determine the type of living cells .

Hereinafter, a method of measuring the three-dimensional refractive index distribution of a cell through optical measurement will be described in more detail.

3A is a view for explaining a method of measuring a three-dimensional refractive index of a cell using an incident light rotation method according to an embodiment.

All objects have a refractive index distribution. The refractive index is the intrinsic optical quantity of the material itself, which describes how much the velocity slows down as the light passes through it. Optical diffraction tomography or optical projection tomography (tomographic phase microscopy, 3D digital holographic microscopy) can be used to measure the three-dimensional refractive index of cells [Non-Patent Documents 1 and 2].

Referring to FIG. 3A, the optical diffraction tomography and the optical projection tomography can use the same optical implementation (Non-Patent Document 3). The light emitted from the coherent light source 310 may be incident on the cell 300 and the transmitted light diffracted by the cell 300 may be measured using the interferometer 320. [

At this time, the three-dimensional refractive index profile 340 of the cell 300 can be measured using several pieces of two-dimensional holograms measured while rotating the angle of incidence on the cell 300. Here, the difference between the optical diffraction tomography and the optical projection tomography is in the reconstruction algorithm 330 considering the diffraction of light in the specimen.

On the other hand, the method of measuring the cell 300 may include a method in which cells are placed at low concentration on an in vitro slide glass, a method in which the cells are singly or in a high concentration on an in vitro slide glass, Multi-layered forms, tissue slides in which biopsy slides are cut to a thickness of between 5 micrometers and 50 micrometers, and microfluidic tubes for high-throughput screening in vitro (microfluidic channel).

And the light source 310 used may utilize a single wavelength laser. Further, by measuring the three-dimensional refractive index (340) at each wavelength by using the multiple wavelength laser as the light source 310, a larger amount of information can be utilized for the classification of the cells 300.

FIG. 3B is a view for explaining a method of measuring the three-dimensional refractive index of cells using the cell rotation method according to one embodiment.

Referring to FIG. 3B, in the method of measuring the three-dimensional refractive index of the cell 300 using the incident light rotation method according to the embodiment described with reference to FIG. 3A, instead of rotating the incident light, (340).

FIG. 4 is a diagram for explaining deep learning using a data set according to an embodiment and cell type classification. Referring to FIG.

The three-dimensional refractive index images of the cells already classified by the method shown in FIG. 3 can be measured. At this time, as shown in FIG. 4, a large number of (> 100) specimens may be measured for each type, and a refractive index characteristic specific to the cell type may be extracted 410 using a deep learning algorithm .

More specifically, based on the measured three-dimensional refractive index information, a convolutional neural network algorithm (Non-Patent Document 4) can be used. In this case, the information input to the neural network algorithm is three-dimensional refractive index information of each cell, and the predicted value resulting from the machine learning becomes the kind or species of the cell.

First, a characteristic extraction is performed (step 410) by sequential stepwise application of convolution and subsampling in the three-dimensional refractive index tomogram data 400 of the cell, Followed by a classification step 420 for classifying the species and species of the cells.

In this case, if the refractive index information composed of the three-dimensional shape is n (x, y, z), it can extract the specific property directly in the form of a three-dimensional matrix. z) can be transformed into a two-dimensional or one-dimensional vector form such as n (k, l) or n (m), and specific properties can be extracted. The position space location (x, y, z) can be transformed into the spatial frequency location (kx, ky, kz) by using Fourier transformation or the like.

By using the specific characteristics of the cell type obtained by the above method, the kind of the cell can be grasped by applying it to the three-dimensional refractive index distribution of a specific cell.

At this time, it is possible to increase the efficiency of the deep run by using both the real part of the refractive index related to the deceleration of light in the material and the imaginary part of the refractive index related to the absorption of light.

FIGS. 5 to 7 are flowcharts for explaining an unoccluded unlabeled cell sorting method using a three-dimensional refractive index tomogram and dip learning according to an embodiment.

Referring to FIG. 5, a method of distinguishing un-stained unlabeled cells using a three-dimensional refractive index tomogram and a deep run according to an embodiment includes a step 510 of optically measuring a three-dimensional refractive index distribution of a cell, (520) of applying a deep learning algorithm or a convolutional neural network algorithm using the 2D refractive index distribution to classify the type of cells into an un-stained unlabeled cell.

6, the step of optically measuring the three-dimensional refractive index distribution of the cell includes a step 511 of introducing the light emitted from the light source into the cell, a step 511 of measuring the transmitted light diffracted by the cell using an interferometer, (Step 512) of obtaining a three-dimensional refractive index profile of the cell, and measuring a three-dimensional refractive index distribution of the cell using a plurality of two-dimensional holograms (step 513).

In addition, referring to FIG. 7, the step of sorting the cells may include extracting a refractive index characteristic specific to a cell type through a convolutional neural network algorithm using the measured three-dimensional refractive index distribution Step 521, and a classification step 522 for classifying the cell type and species according to the refractive index characteristic specific to the extracted cell type.

According to the embodiments, the refractive index distribution of the three-dimensional cell is optically measured, and deep learning or convolutional neural network algorithm is applied to determine the type of living cells And a method and apparatus for distinguishing non-stained non-labeled cell types using deep running.

Hereinafter, a non-stained non-labeled cell sorting method using a three-dimensional refractive index tomogram and a deep running according to one embodiment will be described in more detail using one embodiment.

The non-stained non-labeled cell type classification method using the three-dimensional refractive tomogram and the dip running according to one embodiment can be classified into a non-staining non-labeled cell type classification using the three-dimensional refractive tomogram and the deep running according to the embodiment described in FIG. More specifically, it can be explained using a device. The non-stained non-labeled cell sorting apparatus using the three-dimensional refractive index tomogram and the dip running according to one embodiment may include a three-dimensional refractive index measuring unit and a cell type determining unit. Here, the three-dimensional refractive index measuring unit may include a light source, an interferometer, and a measuring unit. The cell type determining unit may include a characteristic extracting unit and a classifying unit.

In step 510, the three-dimensional refractive index measuring unit may optically measure the three-dimensional refractive index distribution of the cell.

More specifically, in step 511, the light source may cause light to enter the cell to optically measure the three-dimensional refractive index profile of the cell. Then, in step 512, the interferometer can acquire a plurality of two-dimensional holograms by measuring the transmitted light diffracted by the cells after the light incident from the light source is incident on the cells. Thereafter, in step 513, the measuring unit can measure the three-dimensional refractive index distribution of the cell using a plurality of two-dimensional holograms obtained in the interferometer.

The three-dimensional refractive index measuring unit can measure the three-dimensional refractive index distribution of the cell through optical measurement of at least one of optical diffraction tomography and optical projection tomography.

The three-dimensional refractive index measuring unit can measure the three-dimensional refractive index distribution of a cell using a plurality of two-dimensional holograms measured in the interferometer by rotating the angle of light incident on the cell.

In addition, the 3D refractive index measuring unit can directly measure the three-dimensional refractive index distribution of the cells using a plurality of two-dimensional holograms measured in the interferometer by rotating the cells directly.

In step 520, the cell type determination unit applies deep learning or a convolutional neural network algorithm using the three-dimensional refractive index measured by the three-dimensional refractive index measuring unit, The types of cells can be distinguished.

In step 521, the characteristic extractor may extract a refractive index characteristic specific to the cell type through a convolutional neural network algorithm using the measured three-dimensional refractive index distribution.

In addition, the characteristic extracting unit may extract a refractive index characteristic specific to a cell type through a depth learning algorithm using the measured three-dimensional refractive index distribution.

In step 522, the classification unit may classify the cell type and species according to the refractive index characteristic specific to the cell type extracted from the characteristic extraction unit.

According to the embodiments, the types of cells can be distinguished without using dyeing or labeling in the cells. This can be widely used in the field of cell biology, particularly in the field of cytometry.

For example, since a stem cell can be identified by a non-labeled non-staining method, it can be injected into the body immediately. As another example, when the present invention is applied to the field of immunotherapy in which a specific type of leukocyte extracted from blood of a cancer patient is selectively cultured and re-infused, the conventional magnetic cell-based cell sorting (MACS, magnetic-assisted cell cytometry can be effectively solved.

It is also applicable to medical infections in medicine. For example, by making it possible to quickly and accurately distinguish between bacterial species, a medical prescription, such as antibiotics, can be made quickly and effectively for infected patients. Another example is the ability to quickly and accurately screen toxic fungi (eg, anthrax) that can be dedicated to weapons, and thus can be used in counter-terrorism applications.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the apparatus and components described in the embodiments may be implemented within a computer system, such as, for example, a processor, controller, arithmetic logic unit (ALU), digital signal processor, microcomputer, field programmable array (FPA) A programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For ease of understanding, the processing apparatus may be described as being used singly, but those skilled in the art will recognize that the processing apparatus may have a plurality of processing elements and / As shown in FIG. For example, the processing apparatus may comprise a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of the foregoing, and may be configured to configure the processing device to operate as desired or to process it collectively or collectively Device can be commanded. The software and / or data may be in the form of any type of machine, component, physical device, virtual equipment, computer storage media, or device , Or may be permanently or temporarily embodied in a transmitted signal wave. The software may be distributed over a networked computer system and stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (10)

Optically measuring a three-dimensional refractive index profile of cells that are unstained and unlabeled; And
Dividing the cell type into non-staining non-labeling by applying a deep learning algorithm or a convolutional neural network algorithm using the measured three-dimensional refractive index distribution
Lt; / RTI >
The step of optically measuring the three-dimensional refractive index profile of the cell comprises:
Introducing light from the light source into the cell;
Measuring a transmitted light diffracted by the cell with an interferometer to obtain a plurality of two-dimensional holograms; And
Measuring a three-dimensional refractive index profile of the cell using the plurality of two-dimensional holograms,
Dimensional ultrasound image by rotating a angle of light incident on the cell using a digital micromirror device to measure a three-dimensional (3D) image of the cell using the plurality of two-dimensional holograms measured in the interferometer, The refractive index distribution is measured,
The step of sorting the cells includes:
The measured three-dimensional refractive index distribution is input to the deep learning algorithm or the convolutive neural network algorithm to extract refractive index characteristics specific to the cell type based on morphological and chemical characteristics of the cell, Classifying the cell type and species according to the refractive index characteristic specific to the cell type
A method of distinguishing non - stained non - labeled cells using three - dimensional refractive tomograms and deep - running. delete delete delete The method according to claim 1,
The step of optically measuring the three-dimensional refractive index profile of the cell comprises:
Measuring the three-dimensional refractive index distribution of the cell through optical measurement of at least one of optical diffraction tomography and optical projection tomography
A method of distinguishing non - stained non - labeled cells using three - dimensional refractive tomograms and deep - running. The method according to claim 1,
The step of sorting the cells includes:
Extracting a refractive index characteristic specific to a cell type through the convolutional neural network algorithm using the measured three-dimensional refractive index distribution; And
A classification step of classifying the cell type and species according to the refractive index characteristic specific to the extracted cell type;
A method of distinguishing non - stained non - labeled cells using three - dimensional refractive tomograms and deep - running. The method according to claim 1,
The step of sorting the cells includes:
Extracting a refractive index characteristic specific to a cell type through the depth learning algorithm using the measured three-dimensional refractive index distribution; And
A classification step of classifying the cell type and species according to the refractive index characteristic specific to the extracted cell type;
A method of distinguishing non - stained non - labeled cells using three - dimensional refractive tomograms and deep - running. A three-dimensional refractive index measuring unit for optically measuring a three-dimensional refractive index distribution of cells that are unstained and unlabeled; And
A cell type determination unit for classifying the type of the cell with an un-stained non-label by applying deep learning or a convolutional neural network algorithm using the measured three-dimensional refractive index distribution,
Lt; / RTI >
Wherein the three-dimensional refractive index measuring unit comprises:
A light source for introducing light into the cell;
An interferometer for measuring transmitted light diffracted in the cell to obtain a plurality of two-dimensional holograms; And
Dimensional hologram, and measuring a three-dimensional refractive index distribution of the cell using the plurality of two-dimensional holograms,
Dimensional ultrasound image by rotating a angle of light incident on the cell using a digital micromirror device to measure a three-dimensional (3D) image of the cell using the plurality of two-dimensional holograms measured in the interferometer, The refractive index distribution is measured,
The cell type determination unit may determine,
The measured three-dimensional refractive index distribution is input to the deep learning algorithm or the convolutive neural network algorithm to extract refractive index characteristics specific to the cell type based on morphological and chemical characteristics of the cell, Classifying the cell type and species according to the refractive index characteristic specific to the cell type
And a non-staining non-labeled cell type discriminating device using a depth learning. 9. The method of claim 8,
Wherein the three-dimensional refractive index measuring unit comprises:
Dimensional refractive index distribution of the cell is measured by optical measurement of at least one of an optical diffraction tomography and an optical projection tomography, and a non-stained non-labeled cell type discrimination device using a deep- . 9. The method of claim 8,
The cell type determination unit may determine,
A characteristic extractor for extracting a refractive index characteristic specific to a cell type through the depth learning algorithm or the convolutional neural network algorithm using the measured three-dimensional refractive index distribution; And
A classification unit for classifying the type of the cell and the species according to the refractive index characteristic specific to the extracted cell type;
And a non-staining non-labeled cell type discriminating device using a depth learning.

Images