KR20220106582A - Method and apparatus for processing pathological tissue image - Google Patents

Abstract

The present specification provides a method and a device for processing a pathological tissue image, which may determine whether an extracted tissue pathological image is blurred, generate an error message when the blurring of the tissue pathological image is lower than the reference value, check whether a foreign substance is included in the tissue pathological image to remove the foreign substance, and standardize the color of the tissue pathological image.

Description

Pathological tissue image processing method and apparatus {METHOD AND APPARATUS FOR PROCESSING PATHOLOGICAL TISSUE IMAGE}

The present specification discloses a method and apparatus for processing pathological tissue images.

The procedure of biopsy has so far consisted of tissue extraction, tissue staining, and diagnosis by a pathologist using a microscope. However, as studies using computer-aided CAD or deep learning techniques have increased recently, the number of cases of digitizing tissue images with an optical scanner after tissue staining is increasing. In this case, there are problems that did not become a problem when a person observed/diagnosed.

For example, there are situations in which the image looks blurry due to out of focus, or an artifact such as dust/hair enters.

In this case, humans recognize and refer to the error, diagnose or fix the problem, and then repeat the procedure again, but the computer does not recognize it and applies it immediately. For this reason, performance may be degraded, or algorithms for diagnosing diseases may show completely different results.

In addition, different chromosomes or scanners used for each institution/hospital are different when staining tissue, which shows different results with poor objectivity for the same tissue.

Embodiments provide a histopathological imaging method and apparatus capable of deriving an objective and consistent result in tissue diagnosis.

The present specification determines whether the extracted histopathological image is blurred, generates an error message when the blur of the histopathological image is lower than a reference value, checks whether the histopathological image contains a foreign substance, removes the foreign material, and the histopathological image Provided are a method and an apparatus for processing a pathological tissue image for standardizing the color of

In one aspect, the method according to an embodiment includes determining whether the extracted histopathological image is blurred, generating an error message when the blurring of the histopathological image is lower than a reference value, and determining whether a foreign material is included in the histopathological image It includes the steps of confirming and removing the foreign body and standardizing the color of the histopathology image.

In another aspect, the apparatus according to another embodiment includes an input unit receiving the extracted histopathological image, determining whether the extracted histopathological image is blurred, and generating an error message when the blurring of the histopathological image is lower than a reference value, and generating an error message, It includes a control unit for checking whether a foreign material is included in the image, removing the foreign material, standardizing the color of the histopathology image, and an output unit for outputting an error message or outputting a standardized histopathology image.

According to the method and apparatus for histopathological imaging according to the embodiments, since a histopathological image with a focus, a foreign substance removed, and a color standardized is provided, an objective and consistent result can be derived at the time of tissue diagnosis.

1 is a flowchart of a method for processing a pathological tissue image according to an exemplary embodiment.
FIG. 2 is a diagram illustrating a process of calculating a Laplacian variance to determine whether a histopathology image flows in step S110 of FIG. 1 .
3 is a flowchart of the step (S120) of confirming whether a foreign material is included in the histopathological image of FIG. 1 and removing the foreign material.
FIG. 4 is a view for explaining a process of detecting an object through a contour map with respect to a histopathological image in S121 of FIG. 3 .
FIG. 5 is a configuration diagram of a deep learning model used to confirm whether foreign substances are included in the histopathology image in FIG. 3 .
FIG. 6 is a diagram illustrating a process of converting a foreign material into the same color as the background when it is determined that the foreign material is located in the background in step S125.
FIG. 7 is a diagram illustrating a process of standardizing a color or color of a histopathological image by applying a stain/color normalization algorithm in step S130 of FIG. 1 .
8 is a block diagram of an apparatus for processing a pathological tissue image according to another exemplary embodiment.
9 is a diagram illustrating a pathological tissue image processing system for executing a method for processing a pathological tissue image according to an exemplary embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily carry out the present invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is "connected" to another part, this includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween. . Also, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated, and one or more other features However, it is to be understood that the existence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded in advance.

The terms "about," "substantially," and the like, to the extent used throughout the specification are used in or close to the numerical values when manufacturing and material tolerances inherent in the stated meaning are presented, and are intended to enhance the understanding of the present invention. To help, precise or absolute figures are used to prevent unfair use by unscrupulous infringers of the stated disclosure. As used throughout the specification of the present invention, the term "step of (to)" or "step of" does not mean "step for".

In this specification, a "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be implemented using two or more hardware, and two or more units may be implemented by one hardware.

In this specification, some of the operations or functions described as being performed by the terminal, apparatus, or device may be performed instead of in a server connected to the terminal, apparatus, or device. Similarly, some of the operations or functions described as being performed by the server may also be performed in a terminal, apparatus, or device connected to the server.

In this specification, some of the operations or functions described as mapping or matching with the terminal means mapping or matching the terminal's unique number or personal identification information, which is the identification data of the terminal. can be interpreted as

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a flowchart of a method for processing a pathological tissue image according to an exemplary embodiment.

Referring to FIG. 1 , the method 100 for processing a pathological tissue image according to an embodiment determines whether an extracted histopathological image is blurred, checks whether a foreign material is included in the histopathological image, and the color of the histopathological image standardize

Specifically, the method 100 for processing a pathological tissue image according to an embodiment determines whether the extracted histopathological image is blurred, and generates an error message when the blur of the histopathological image is lower than a reference value (S110), It is confirmed whether a foreign material is included in the histopathological image, and includes the step of removing the foreign material (S120) and the step of standardizing the color of the histopathological image (S130). However, step S130 may be performed before and after steps S110 and S120, or may be omitted.

FIG. 2 is a view for explaining a process of determining whether a histopathology image flows with respect to a digital image of histopathology in step S110 of FIG. 1 .

Referring to FIG. 2 , when determining whether the histopathology image is blurred, the extracted tissue pathology is first stained and then scanned or photographed to be digitized.

There are two methods for digitizing histopathology: a snapshot method that takes a picture of a certain area under light magnification conditions of a constant magnification, and a whole slide imaging (WSI) method that digitizes some or all of the contents of a glass slide. , the latter is often referred to as a virtual microscope. A camera may be used for the instantaneous shooting method, and a digital scanner may be used for imaging the entire slide.

The present specification exemplarily describes, but is not limited to, imaging the entire slide using a digital scanner as a method of digitizing histopathology.

Thereafter, various methods or algorithms for determining whether the histopathology image is blurred may be used in order to check whether the stored image is in focus with respect to the histopathology digital image.

For example, a variance of Laplacian may be calculated for the histopathology digital image.

As shown in FIG. 2 , the Laplacian variance can be calculated by using the kernel matrix as in Equation 1 for the histopathological image.

[Equation 1]

An error message is generated when the Laplacian variance is lower than the reference value, which varies according to the size of the histopathology digital image. For example, when the Laplacian variance is lower than a reference value that varies according to the size of the histopathology digital image, a rescan or re-scan may be requested through a generated error message.

If the Laplacian variance is equal to or greater than the reference value that varies according to the size of the histopathology digital image, proceed to the next step.

3 is a flowchart of the step (S120) of confirming whether a foreign material is included in the histopathological image of FIG. 1 and removing the foreign material.

Referring to FIG. 3 , the step of confirming whether a foreign material is included in the histopathological image and removing the foreign material (S120) is to detect the object in the histopathological image (S121), confirm the foreign material (S122), and the foreign material to the tissue It is determined whether it overlaps with the tissue (S123), and when it is determined that the foreign material overlaps the tissue, a rescan or re-photography is requested (S124), and when it is determined that the foreign material does not overlap the tissue, the foreign material is removed (S125).

When the object is detected in S121 , the object may be detected through a contour map as shown in FIG. 4 for the histopathological image.

When it is checked whether a foreign material is included in the histopathology image in S122, it may be checked whether an artifact such as hair or dust is included in the histopathological image using artificial intelligence.

FIG. 5 is a configuration diagram of a deep learning model used to determine whether foreign substances are included in the histopathology image in FIG. 3 .

Referring to FIG. 5 , artificial intelligence uses a deep learning model 221 in a deep neural network consisting of a multi-layered network to learn a large amount of histopathological images by learning the characteristics of the input value ( feature), and through this, a multi-layered network is trained to minimize the error in prediction accuracy, so that it is possible to check whether foreign substances are included in the histopathology image.

The deep learning model 221 used to check whether a foreign material is included in the histopathology image may be a model in which an artificial neural network is stacked in multiple layers. That is, the deep learning model automatically learns features of the input value by learning a large amount of data from a deep neural network consisting of a multi-layered network, and through this, the error in the objective function, that is, the prediction accuracy, is eliminated. It is a type of model that trains the network to minimize it.

The above-described deep learning model 221 is a Convolutional Neural Network (CNN), a Deep Hierachical Network (DHN), a Convolutional Deep Belief Network (CDBN), a Deconvolutional Deep Network (DDN), a Recurrent Neural Network (RNN), and a Generative Adversarial Network (GAN). ), but the present invention is not limited thereto, and various deep learning models that can be used now or in the future can be used.

The above-described deep learning model 221 may be implemented through a deep learning framework. The deep learning framework provides functions commonly used when developing the deep learning model 221 in the form of a library, and serves to support good use of system software or hardware platforms. In the present embodiment, the deep learning model 221 may be implemented using any deep learning framework that is currently public or will be released in the future.

Referring back to FIG. 5, the deep learning model 221 performs convolution and subsampling on the input histopathological image to extract a picture for the histopathological image (feature). a feature extraction part 222) and an output part 224 for outputting whether the foreign material is included in the histopathology image using the extracted picture.

Convolution creates a feature map by using a plurality of filters for each region of a medical image in a convolution layer. Subsampling or polling reduces the size of a feature map in a subsampling layer to extract a feature for an image that is invariant to a change in position or rotation.

The picture extraction region 222 extracts pictures of various levels from low-level pictures such as points, lines, and planes to complex and meaningful high-level pictures in a medical image by repeating convolution and/or subsampling. can do.

A deep learning model, for example, a CNN-based deep learning model, aims to optimally learn parameters present in each individual layer in the picture extraction region 222 and the output region 224 . In a deep learning model, the order of the data determines the value of the initial parameter.

The deep learning model 221 may apply random sampling (data order random) and a regulation technique. Random sampling means that the order of the training data learned from the training data set is different.

Regulatory technology is a technology that reduces overfitting, in which a deep learning model over-trained on training data including noise is less accurate in testing or diagnosis. The regulatory technology may be, for example, a drop-out technology or a drop connected technology.

The dropout technique is a method of probabilistically assigning a parameter value to 0 for a specific node and learning it. Drop connected technology is a method of dropping and learning connections between nodes. The dropout technique is exemplarily described below as a regulatory technique below, but it may be any technique or algorithm that reduces overfitting now or in the future.

Referring back to FIG. 3 , in S123 , when it is confirmed that foreign substances are included in the histopathology image, an image is generated in which a part containing tissue in the histopathology image is expressed as a binary mask, and the foreign substances are separated from the tissue. You can figure out whether they are overlapping or located in the background.

If it is determined in S124 that the foreign material overlaps the tissue, a rescan or re-photography may be requested through an error message, and the first step may be restarted.

In S125 , when it is determined that the foreign material is located in the background, as shown in FIG. 6 , the foreign material may be converted into the same color as the background.

According to steps S110 and S120, if the scan result is out of focus or there is a foreign substance, so that the most objective and consistent result can be derived when the CAD or deep learning model is applied after scanning the histopathology image after staining. It can be removed from the existing histopathology image, or an error message can be displayed to induce re-capturing or rescan.

Referring back to FIG. 1 , in the step of standardizing the color of the histopathology image ( S130 ), in order to consistently convert equipment and chromosomes used in institutions/hospitals, as shown in FIG. 7 , color normalization (stain/color normalization) An algorithm can be applied to standardize the color or color of the histopathology image.

By step S130, since a stain/color normalization algorithm is also applied, the consistency of histopathology images can be maintained regardless of equipment and chromosomes used in institutions/hospitals.

According to the method of histopathological imaging according to the above-described embodiments, since a histopathological image with a focus, a foreign substance removed, and a color standardized is provided, an objective and consistent result can be derived at the time of tissue diagnosis.

The method 100 for processing a pathological tissue image according to the above-described embodiment determines whether the extracted histopathological image is blurred, and generates an error message when the blur of the histopathological image is lower than a reference value (S110), tissue Checking whether a foreign material is included in the pathological image, removing the foreign material (S120) and standardizing the color of the histopathological image (S130) are included.

However, the method 100 for processing a pathological tissue image according to an embodiment not only includes all of steps S110, S120, and S130, but also includes only one of steps S110, S120, and S130, or only two of them. may be

For example, the method 100 for processing a pathological tissue image according to an embodiment may consist of only step S120 described with reference to FIG. 3 .

A method of processing a pathological tissue image according to an embodiment has been described above with reference to FIGS. 1 to 7 . Hereinafter, an apparatus for processing a pathological tissue image according to another exemplary embodiment will be described with reference to FIG. 8 .

8 is a block diagram of an apparatus for processing a pathological tissue image according to another exemplary embodiment.

The apparatus 200 for processing a pathological tissue image according to another embodiment includes the input unit 210 for receiving the extracted histopathological image, and determines whether the extracted histopathological image is blurred, and when the blur of the histopathological image is lower than a reference value A control unit ( S220 ) that generates an error message, checks whether a foreign material is included in the histopathological image, removes the foreign material, and standardizes the color of the histopathological image and outputs an error message or a standardized histopathological image and an output unit (S230).

When determining whether the histopathology image is blurred, the control unit 220 digitizes the extracted tissue pathology by scanning or photographing after staining, calculating the variance of Laplacian for the histopathology digital image, and When the Laplacian variance is lower than a reference value that varies according to the size of the histopathology digital image, an error message may be generated.

When determining whether a foreign material is included in the histopathological image, the controller 220 may detect an object through a contour map with respect to the histopathological image.

When determining whether a foreign material is included in the histopathology image, the controller 220 may determine whether the foreign material is included in the histopathological image using artificial intelligence.

As described above, artificial intelligence applies the deep learning model 221 in a deep neural network consisting of a multi-layered network to learn a large amount of tissue pathology images to automatically learn features of the input values. It is possible to learn whether the foreign material is included in the histopathology image by learning the multi-layered network to minimize the error in prediction accuracy through learning.

As described above with reference to FIG. 5, the deep learning model 221 performs convolution and subsampling on the input histopathological image to extract a picture for the histopathological image (feature). It may include a picture extraction area 222 and an output area 224 for outputting whether the foreign material is included in the histopathology image using the extracted picture.

When it is confirmed that a foreign material is included in the histopathology image, the control unit 220 generates an image expressing a part containing the tissue in the histopathological image as a binary mask to determine whether the foreign material overlaps the tissue or is located in the background can

When it is determined that the foreign material is located in the background, the controller 220 may convert the foreign material to the same color as the background.

The controller 220 continuously learns the artificial intelligence using the continuously collected learning image image and the accumulated chart data so that the artificial intelligence can read the image image.

The display unit 230 displays information on the treatment site when a specific treatment area of the image image is selected. When the image of the treatment area is selected, a treatment summary memo window is displayed, and when the treatment summary memo window is selected, the corresponding treatment area is selected. A chart for the treatment area is printed.

As described above with reference to FIG. 8 , the deep learning model 221 that can be used in the controller 220 may be a model in which an artificial neural network is stacked in multiple layers.

As described above with reference to FIG. 8 , the deep learning model 221 includes a picture extraction area 222 and an output area 224 . A deep learning model, for example, a CNN-based deep learning model, aims to optimally learn parameters present in each individual layer in the picture extraction region 222 and the output region 224 . The deep learning model 221 may apply random sampling (data order random) and a regulation technique. The deep learning models 221 use highly flexible nonlinear algorithms.

According to the apparatus for imaging histopathology according to another embodiment described above, it is possible to derive an objective and consistent result at the time of tissue diagnosis.

An apparatus for processing a pathological tissue image according to another embodiment has been described above with reference to FIG. 8 . Hereinafter, a pathological tissue image processing system for executing a pathological tissue image processing method according to an exemplary embodiment will be described with reference to FIG. 9 .

9 illustrates a pathological tissue image processing system for executing a method for processing a pathological tissue image according to an exemplary embodiment.

Referring to FIG. 9 , the pathological tissue image processing system 300 in which the method 100 for processing a pathological tissue image shown in FIG. 1 is executed is connected to the scanner 310 by a scanner 310 and a cable 330 and and a computer device 320 for performing the method 100 for processing the pathological tissue image and displaying the result. The computer device 320 may be a computer device that implements instructions for executing the method 100 for processing a pathological tissue image according to the above-described embodiment, included in a computer-readable recording medium.

In this case, the pathological tissue processing system 300 may be configured by integrating the scanner 310 and the computer device 320 .

The method 100 for processing a pathological tissue image according to the above-described embodiment may be executed by an application basically installed in the computer device 320 , and a user may use an application store server, an application, or a web server related to the corresponding service. It is compatible with the operating system of the computer device 320 through an application providing server, such as, and may be executed by an application (ie, a program) installed directly on the computer device 320 . Here, the operating system of the computer device 320 is an operating system such as Windows or Macintosh installed in a general PC such as a desktop, or iOS, Android ( It may be a mobile-only operating system such as Android).

In this sense, the method 100 for processing a pathological tissue image according to an embodiment is implemented as an application (ie, a program) installed by default in the computer device 320 or directly installed by a user, and is transmitted to the computer device 320 . It may be recorded on a readable recording medium 322 .

The computer program 124 implementing the method 100 for processing a pathological tissue image according to an embodiment corrects the amplification curve obtained from the amplification reaction performed in the presence of the signal generating means according to the embodiment of the present invention described above. All functions corresponding to the method can be executed.

The computer program 324 is a computer program including program instructions stored in a computer-readable recording medium constituting a processor for executing the method 100 for processing a pathological tissue image according to an embodiment.

The computer device 320-readable recording medium 322 is a computer-readable recording medium including instructions for implementing a processor for executing the method 100 for processing a pathological tissue image according to an embodiment.

In this way, in order for the computer of the computer device 320 to read the computer program recorded on the recording medium and execute the method 100 for processing a pathological tissue image according to an embodiment implemented as a program, the above-described program is a computer program. It may include code coded in a computer language such as C, C++, JAVA, Visual basic, VBScript, JavaScript, Perl, XML, or machine language that a processor (CPU) can read.

The above-described method 100 for processing a pathological tissue image may be implemented by a computing device including at least a portion of a processor, a memory, a user input device, and a presentation device. A memory is a medium that stores computer-readable software, applications, program modules, routines, instructions, and/or data, etc. coded to perform specific tasks when executed by a processor. The processor can read and execute computer-readable software, applications, program modules, routines, instructions, and/or data stored in the memory, and the like. The user input device may be a means for allowing the user to input a command to the processor to execute a specific task or to input data required for the execution of the specific task. The user input device may include a physical or virtual keyboard or keypad, key button, mouse, joystick, trackball, touch-sensitive input means, or a microphone. The presentation device may include a display, a printer, a speaker, or a vibrator.

The computing device 320 may include a variety of devices, such as a smartphone, tablet, laptop, desktop, server, client, and the like. The computing device may be a single stand-alone device, or may include a plurality of computing devices operating in a distributed environment comprising a plurality of computing devices cooperating with each other through a communication network.

In addition, the above-described method 100 for processing a pathological tissue image includes a processor, and when executed by the processor, computer-readable software, application, and program module coded to perform an image diagnosis method using a deep learning model , routines, instructions, and/or data structures may be executed by a computing device having a memory stored therein.

The above-described embodiments may be implemented through various means. For example, the present embodiments may be implemented by hardware, firmware, software, or a combination thereof.

In the case of hardware implementation, the method 100 for processing a pathological tissue image using a deep learning model according to the present embodiments includes one or more ASICs (Application Specific Integrated Circuits), DSPs (Digital Signal Processors), DSPDs (Digital Signal Processing Devices), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), processors, controllers, microcontrollers, or microprocessors may be implemented.

For example, the method 100 for processing a pathological tissue image according to embodiments may be implemented using an artificial intelligence semiconductor device in which neurons and synapses of a deep neural network are implemented with semiconductor elements. In this case, the semiconductor device may be a currently used semiconductor device, for example, SRAM, DRAM, NAND, or the like, or a next-generation semiconductor device, RRAM, STT MRAM, PRAM, or the like, or a combination thereof.

When the method 100 for processing a pathological tissue image according to the embodiments is implemented using an artificial intelligence semiconductor device, the result (weight) of learning a deep learning model with software is transferred to a synaptic mimic device arranged in an array, or an artificial intelligence semiconductor Learning can also take place on the device.

In the case of implementation by firmware or software, the method 100 for processing a pathological tissue image according to the present embodiments may be implemented in the form of an apparatus, procedure, or function for performing the functions or operations described above. The software code may be stored in the memory unit and driven by the processor. The memory unit may be located inside or outside the processor, and may transmit and receive data to and from the processor by various known means.

Also, as described above, terms such as "system", "processor", "controller", "component", "module", "interface", "model", or "unit" generally refer to computer-related entities hardware, hardware and software. may mean a combination of, software, or running software. For example, the aforementioned components may be, but are not limited to, a process run by a processor, a processor, a controller, a controlling processor, an object, a thread of execution, a program, and/or a computer. For example, both an application running on a controller or processor and a controller or processor can be a component. One or more components may reside within a process and/or thread of execution, and the components may be located on one device (eg, a system, computing device, etc.) or distributed across two or more devices.

And, in consideration of the system environment of a computer that reads a recording medium and executes a program by reading a recording medium, a functional program and related codes and code segments for implementing the present invention, programmers in the technical field to which the present invention belongs may be easily inferred or changed by

The method 100 for processing a pathological tissue image may also be implemented in the form of a recording medium including instructions executable by a computer, such as an application or program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer-readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The above-described method 100 for processing a pathological tissue image may be executed by an application that is basically installed in the terminal (which may include a program included in a platform or an operating system, etc. that is basically installed in the terminal), and the user selects the application store It may be executed by an application (ie, a program) installed directly on the master terminal through an application providing server such as a server, an application, or a web server related to the corresponding service. In this sense, the above-described glaucoma surgery result diagnosis method may be implemented as an application (ie, a program) installed basically in a terminal or directly installed by a user, and may be recorded in a computer-readable recording medium such as a terminal. The above description of the present invention is for illustration, and those of ordinary skill in the art to which the present invention pertains can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a dispersed form, and likewise components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be interpreted as being included in the scope of the present invention. do.

Claims (16)

determining whether the extracted histopathology image is blurred, and generating an error message when the blurring of the histopathology image is lower than a reference value;
determining whether a foreign material is included in the histopathological image, and removing the foreign material; and
and normalizing the color of the histopathology image. According to claim 1,
When determining whether the histopathological image is blurred,
After staining the extracted histopathology, it is scanned or photographed and digitized, and a variance of Laplacian is calculated for the histopathology digital image. How to generate the above error message if low. The method of claim 1,
A method of detecting an object through a contour map with respect to the histopathological image when it is determined whether a foreign material is included in the histopathological image. The method of claim 1,
When determining whether a foreign material is included in the histopathological image, a method of confirming whether the foreign material is included in the histopathological image using artificial intelligence. 5. The method of claim 4,
The artificial intelligence applies a deep learning model in a deep neural network consisting of a multi-layered network to learn a large amount of tissue pathology images to automatically learn features of the input value, and predict accuracy through this A method of determining whether the foreign material is included in the histopathology image by training the multi-layered network to minimize the error of 6. The method of claim 5,
The deep learning model performs convolution and subsampling on the input histopathological image to extract a picture for the histopathological image (feature extraction part); and an output part for outputting whether the foreign material is included in the histopathology image by using a picture. According to claim 1,
When it is confirmed that foreign substances are included in the histopathology image, an image is generated in which a portion containing tissue in the histopathology image is expressed as a binary mask, and the method of determining whether the foreign material overlaps the tissue or is located in the background . 8. The method of claim 7,
When it is determined that the foreign material is located in the background, the method of converting the foreign material to the same color as the background. an input unit for receiving the extracted histopathology image;
It is determined whether the extracted histopathology image is blurred, and when the blur of the histopathology image is lower than a reference value, an error message is generated, it is checked whether foreign substances are included in the histopathology image, the foreign material is removed, and the histopathology image a control unit for standardizing the color of and
and an output unit for outputting the error message or for outputting the standardized histopathology image. 10. The method of claim 9,
When determining whether the histopathological image is blurred,
After staining the extracted histopathology, it is scanned or photographed and digitized, and a variance of Laplacian is calculated for the histopathology digital image. If low, the device generates the above error message. 10. The method of claim 9,
An apparatus for detecting an object through a contour map with respect to the histopathological image when it is determined whether a foreign material is included in the histopathological image. 10. The method of claim 9,
When determining whether a foreign material is included in the histopathological image, an apparatus for determining whether the foreign material is included in the histopathological image using artificial intelligence. 13. The method of claim 12,
The artificial intelligence applies a deep learning model in a deep neural network consisting of a multi-layered network to learn a large amount of tissue pathology images to automatically learn features of the input value, and predict accuracy through this An apparatus for determining whether the foreign material is included in the histopathology image by learning the multi-layered network to minimize the error of 14. The method of claim 13,
The deep learning model performs convolution and subsampling on the input histopathological image to extract a picture for the histopathological image (feature extraction part); and an output part for outputting whether the foreign material is included in the histopathology image by using a picture. 12. The method of claim 11,
When it is confirmed that foreign substances are included in the histopathological image, an image is generated in which a portion containing tissue in the histopathological image is expressed as a binary mask to determine whether the foreign material overlaps the tissue or is located in the background . 16. The method of claim 15,
When it is determined that the foreign material is located in the background, the apparatus converts the foreign material into the same color as the background.

Images