KR20220082456A - Method for detecting state of object of interest based on medical image - Google Patents

Abstract

According to an embodiment of the present disclosure, a method for predicting a disease based on a medical image performed by a computing device is disclosed. The method includes the steps of generating a feature vector including predictive values of brain disease for each of the 2D medical images included in the 3D medical image, using the pre-trained first model; estimating an importance indicating prediction accuracy for each of the 2D medical images based on the feature vector using a second model, and the brain disease among the 2D medical images based on the importance and selecting at least one model input image suitable for prediction of .

Description

A method of reading the status of a target of interest based on a medical image {METHOD FOR DETECTING STATE OF OBJECT OF INTEREST BASED ON MEDICAL IMAGE}

The present invention relates to a method of processing a medical image, and more particularly, to a method of determining a state of an object of interest in a medical image using a neural network.

Medical images are data that enable us to understand the physical state of various organs of the human body. Medical images include digital radiographic imaging (X-ray), computed tomography (CT), or magnetic resonance imaging (MRI).

Various deep learning technologies have been developed to identify the state of a target of interest among body organs included in a medical image. In particular, due to the nature of the domain with few data resources, interest in deep learning technology for judging the state of the entire object of interest based on information on a part of the object of interest has been steadily increasing. As a representative prior art, there is a technique for reading a state of an object of interest through a deep learning model that receives medical images including a part of the object of interest as one input. This technology reads the entire state of the object of interest by making one input through a method of stacking several medical images including a part of the object of interest into channels (channel concatenation) or a method of concatenating images (spatial concatenation). .

Since the aforementioned representative prior art uses one deep learning model, there are advantages in that the resource required for reading the state of interest is small, and the learning and reading time of the model can be fast. In addition, the prior art has an advantage in that associations between images can be learned. However, in the prior art, since the input image dimension is large and the number of parameters of the deep learning model is relatively small, deep learning learning may be difficult, and there is a problem that the reading performance of the deep learning model may be deteriorated. In addition, since the deep learning model of the prior art uses all features unnecessary for learning and reading, there is a disadvantage that efficient and accurate reading cannot be performed. There is a demand in the art to solve these shortcomings of the prior art.

US Patent Publication No. 2020-0211187 (2020.07.02) discloses a system and method for ossification center detection and bone age evaluation.

The present disclosure has been made in response to the above-described background technology, and an object of the present disclosure is to provide an artificial intelligence algorithm optimized for reading the state of an object of interest included in a medical image.

Disclosed is a medical image-based disease prediction method performed by a computing device according to an embodiment of the present disclosure for realizing the above-described problem. The method includes: extracting first feature vectors from each of a plurality of input images including at least a portion of an object of interest, using a pre-trained reading model; generating a second feature vector based on a correlation between first feature vectors corresponding to each of the plurality of input images by using the read model; and combining each of the first feature vectors with the second feature vector using the read model, and determining the state of the interest based on the combined feature vectors.

In an alternative embodiment, the first feature vectors may each include a local feature vector for the object of interest generated based on at least a part of the object of interest. In addition, the second feature vector may include a global feature vector for the object of interest generated based on a combination of at least some of the first feature vectors.

In an alternative embodiment, the plurality of input images may include: first sub-input images including at least a portion of an object of interest photographed based on a first direction; and second sub-input images including at least a portion of the object of interest photographed based on a second direction different from the first direction.

In an alternative embodiment, the first sub-input images may include front images of at least a portion of a body region corresponding to the object of interest. In addition, the second sub-input images may include side images of at least a portion of the body region.

In an alternative embodiment, the generating of the second feature vector may include determining, using the read model, a correlation between the first feature vectors, and based on a result of the determination, a first correlation exists 1 Selecting feature vectors; and using the read model to generate a second feature vector based on the first feature vectors in which the selected correlation exists.

In an alternative embodiment, the step of determining the state of the object of interest comprises, using the read model, a score for determining the state of the object of interest based on each of the first feature vectors combined with the second feature vector. deriving them; and summing scores corresponding to each of the first feature vectors, and reflecting the result of the summation in biomarker data related to the subject of interest to read the state of the object of interest.

In an alternative embodiment, the scores for determining the status of the target of interest may include scores representing bone maturity of at least a portion of the body region corresponding to the target of interest, and the status of the target of interest includes: It may be the bone age of the body region corresponding to the target of interest.

In an alternative embodiment, the read model may include: a first neural network that receives each of the plurality of input images and generates first feature vectors related to at least a portion of the object of interest; a second neural network that receives a combined feature vector based on the correlation between the first feature vectors and generates a second feature vector related to the object of interest; and a third neural network that receives each of the first feature vectors to which the second feature vector is reflected and generates data about the state of the interest.

In an alternative embodiment, the method includes, using a pre-trained pre-processing model, a center point of a bounding box containing at least a portion of the object of interest based on at least one medical image containing the object of interest. detecting; and generating a plurality of input images including a predetermined region determined based on the central point from the at least one medical image by using the preprocessing model.

Further, in an alternative embodiment, the method includes using the pre-processing model to detect a plurality of reference points for geometric deformation of the object of interest based on a region in which the object of interest exists in the at least one medical image. step; and normalizing the at least one medical image based on the plurality of reference points by using the preprocessing model.

A computer program stored in a computer-readable storage medium is disclosed according to an embodiment of the present disclosure for realizing the above-described problems. When the computer program is executed on one or more processors, the computer program performs the following operations for reading a condition of a subject of interest based on a medical image, the operations comprising: using a pre-trained reading model, extracting first feature vectors from each of a plurality of input images including a portion; generating a second feature vector based on a correlation between first feature vectors corresponding to each of the plurality of input images by using the read model; and combining each of the first feature vectors with the second feature vector using the read model, and determining the state of the interest based on the combined feature vectors.

According to an embodiment of the present disclosure for realizing the above-described problem, a computing device for reading a state of an object of interest based on a medical image is disclosed. The apparatus includes: a processor including at least one core; a memory containing program codes executable by the processor; and a network unit for receiving a medical image, wherein the processor extracts first feature vectors from each of a plurality of input images including at least a portion of an interest by using a pre-trained reading model, and the reading Using a model, a second feature vector is generated based on a correlation between first feature vectors corresponding to each of the plurality of input images, and each of the first feature vectors is calculated by using the read model. The second feature vector may be combined, and the state of the target of interest may be determined based on the combined feature vectors.

The present disclosure may provide an artificial intelligence algorithm optimized for reading a state of a target of interest included in a medical image.

1 is a block diagram of a computing device for reading a state of an object of interest based on a medical image according to an embodiment of the present disclosure.
2 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure;
3 is a block diagram illustrating an operation of a neural network model for reading a state of an object of interest according to an embodiment of the present disclosure.
4 is a conceptual diagram illustrating an architecture of a neural network model for reading a state of an object of interest according to an embodiment of the present disclosure.
5 is a block diagram illustrating a process of reading a state of an object of interest by a computing device according to an embodiment of the present disclosure.
6 is a conceptual diagram illustrating an operation and configuration of a preprocessing model of a computing device according to an embodiment of the present disclosure.
7 is a flowchart of a method for reading a state of an object of interest based on a medical image according to an embodiment of the present disclosure.
8 is a flowchart of a pre-processing method of a medical image according to an embodiment of the present disclosure.
9 is a schematic diagram of a computing environment according to an embodiment of the present disclosure;

Various embodiments are now described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the present disclosure. However, it is apparent that these embodiments may be practiced without these specific descriptions.

The terms “component,” “module,” “system,” and the like, as used herein, refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or execution of software. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, a thread of execution, a program, and/or a computer. For example, both an application running on a computing device and the computing device may be a component. One or more components may reside within a processor and/or thread of execution. A component may be localized within one computer. A component may be distributed between two or more computers. In addition, these components can execute from various computer readable media having various data structures stored therein. Components may communicate via a network such as the Internet with another system, for example, via a signal having one or more data packets (eg, data and/or signals from one component interacting with another component in a local system, distributed system, etc.) may communicate via local and/or remote processes depending on the data being transmitted).

In addition, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from context, "X employs A or B" is intended to mean one of the natural implicit substitutions. That is, X employs A; X employs B; or when X employs both A and B, "X employs A or B" may apply to either of these cases. It should also be understood that the term “and/or” as used herein refers to and includes all possible combinations of one or more of the listed related items.

Also, the terms "comprises" and/or "comprising" should be understood to mean that the feature and/or element in question is present. However, it should be understood that the terms "comprises" and/or "comprising" do not exclude the presence or addition of one or more other features, elements and/or groups thereof. Also, unless otherwise specified or unless it is clear from context to refer to a singular form, the singular in the specification and claims should generally be construed to mean “one or more”.

And, the term "at least one of A or B" should be interpreted as meaning "when including only A", "when including only B", and "when combined with the configuration of A and B".

Those skilled in the art will further appreciate that the various illustrative logical blocks, configurations, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented in electronic hardware, computer software, or combinations of both. It should be recognized that they can be implemented with To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

Descriptions of the presented embodiments are provided to enable those skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art of the present disclosure. The generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Therefore, the present invention is not limited to the embodiments presented herein. This invention is to be interpreted in its widest scope consistent with the principles and novel features presented herein.

In the present disclosure, a network function, an artificial neural network, and a neural network may be used interchangeably.

On the other hand, the term "image" or "image data" as used throughout the detailed description and claims of the present invention refers to multidimensional data composed of discrete image elements (eg, pixels in a two-dimensional image), in other words , a term that refers to a visible object (eg, displayed on a video screen) or a digital representation of that object (eg, a file corresponding to a pixel output of a CT, MRI detector, etc.).

For example, "image" or "image" can be defined by computed tomography (CT), magnetic resonance imaging (MRI), ultrasound or any other medical imaging system known in the art. It may be a medical image of a collected subject. The image does not necessarily have to be provided in a medical context, but may be provided in a non-medical context, for example, there may be X-ray imaging for security screening.

Throughout the detailed description and claims of the present invention, the 'DICOM (Digital Imaging and Communications in Medicine)' standard is a generic term for various standards used for digital image expression and communication in medical devices. The DICOM standard is published by a joint committee formed by the American Society of Radiological Medicine (ACR) and the National Electrical Engineers Association (NEMA).

In addition, throughout the detailed description and claims of the present invention, 'Picture Archiving and Communication System (PACS)' is a term that refers to a system that stores, processes, and transmits in accordance with the DICOM standard, X-ray, CT , medical imaging images acquired using digital medical imaging equipment such as MRI are stored in DICOM format and transmitted to terminals inside and outside the hospital through a network, to which reading results and medical records can be added.

1 is a block diagram of a computing device for reading a state of an object of interest based on a medical image according to an embodiment of the present disclosure.

The configuration of the computing device 100 shown in FIG. 1 is only a simplified example. In an embodiment of the present disclosure, the computing device 100 may include other components for performing the computing environment of the computing device 100 , and only some of the disclosed components may configure the computing device 100 .

The computing device 100 may include a processor 110 , a memory 130 , and a network unit 150 .

The processor 110 may include one or more cores, and a central processing unit (CPU) of a computing device, a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU). unit) and the like, and may include a processor for data analysis and deep learning. The processor 110 may read a computer program stored in the memory 130 to perform data processing for machine learning according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor 110 may perform an operation for learning the neural network. The processor 110 for learning of the neural network, such as processing input data for learning in deep learning (DL), extracting features from the input data, calculating an error, updating the weight of the neural network using backpropagation calculations can be performed. At least one of a CPU, a GPGPU, and a TPU of the processor 110 may process learning of a network function. For example, the CPU and the GPGPU can process learning of a network function and data classification using the network function. Also, in an embodiment of the present disclosure, learning of a network function and data classification using the network function may be processed by using the processors of a plurality of computing devices together. In addition, the computer program executed in the computing device according to an embodiment of the present disclosure may be a CPU, GPGPU or TPU executable program.

According to an embodiment of the present disclosure, the processor 110 may read the state of the object of interest included in the medical image by using the pre-trained neural network model. In the present disclosure, an object of interest may include all body organs such as the brain and lungs, as well as detailed tissues of body components such as bones. In addition, in the present disclosure, the state of interest may refer to a clinical state based on physicochemical changes such as deformation or degeneration of body tissues. The state of the object of interest may vary depending on the type of the object of interest. For example, when the object of interest is a body organ such as brain or lung, the state of interest means a normal state in which a specific disease does not exist in a body organ, an abnormal state in which treatment is required due to the presence of a specific disease in a body organ, etc. can do. When the object of interest is a body composition such as bone, the state of the object of interest may mean a bone age based on a growth degree of a body tissue such as bone.

The processor 110 may convert the medical image into various forms in order to improve the performance of the neural network model for reading the state of the object of interest. The processor 110 may generate input images of the neural network model including at least a portion of the object of interest based on the transformed image. The processor 110 may input a plurality of images generated from the medical image to the neural network model to generate basic data for determining the state of the object of interest. The processor 110 may generate read data indicating the state of the object of interest by using the basic data output through the neural network model and the biomarker data related to the subject of interest.

For example, the processor 110 may read the bone age of the elbow included in a medical image obtained by photographing an upper limb using a pre-trained neural network model. First, the processor 110 may perform geometric transformation of the medical image to improve the performance of the neural network model. The processor 110 may generate a plurality of model input images including elbow regions of interest in the medical image on which geometric deformation is completed. The processor 110 may input a plurality of input images to the neural network model to calculate scores indicating the bone maturity of the elbow regions. The processor 110 may generate a bone age reading result based on the bone maturity of the elbow based on the scores output through the neural network model and biomarker data such as gender.

In detail, the processor 110 may extract first feature vectors from each of a plurality of input images including at least a portion of an interest by using a pre-trained neural network model (hereinafter, a read model). In this case, the first feature vectors may each include a local feature vector for the object of interest generated based on at least a part of the object of interest. For example, the processor 110 may input a plurality of input images including a detailed region of the ROI, which is at least a part of the ROI, to the reading model. The processor 110 may generate first feature vectors that are independent features of each of the plurality of input images by using the read model. In other words, the processor 110 may generate mutually independent first feature vectors that are features corresponding to each of the detailed regions constituting the ROI by using the read model. Accordingly, the first feature vectors generated based on at least a portion of the mutually distinct object of interest (i.e. detailed regions of the object of interest) may indicate a regional feature of the object of interest.

The processor 110 may generate a second feature vector based on a correlation between the first feature vectors corresponding to each of the plurality of input images by using the read model. In this case, the second feature vector may include a global feature vector for the object of interest generated based on a combination of at least some of the first feature vectors. For example, the processor 110 may determine a correlation between the first feature vectors. In addition, the processor 110 may identify first feature vectors that are correlated with each other based on the result of the determination. The processor 110 may combine the first feature vectors in which there is a correlation (i.e. correlation). The processor 110 may generate a second feature vector based on a result of combining the first feature vectors that are correlated using the read model. In other words, the processor 110 may generate a second feature vector including features shared by the detailed regions constituting the ROI by using the read model based on the first feature vectors with the correlation. Accordingly, the second feature vector generated based on a combination of a portion of the first feature vectors may represent a global feature including features shared among detailed regions of interest.

The processor 110 may determine the state of the interest based on the first feature vectors combined with the second feature vector using the read model. For example, the processor 110 may combine the second feature vector and each of the first feature vectors. In other words, the second feature vector representing the global feature of the object of interest may be reflected in each of the first feature vectors representing the local feature of the object of interest. The processor 110 may use the read model to determine the state of the object of interest based on the first feature vectors to which the second feature vector is reflected. The processor 110 may use the read model to determine the state of the object of interest based on mutually independent feature vectors in which correlations between detailed regions of the object of interest are reflected. The first feature vectors to which the second feature vector is reflected help to more accurately and precisely read the state of the object of interest.

According to an embodiment of the present disclosure, the memory 130 may store any type of information generated or determined by the processor 110 and any type of information received by the network unit 150 .

According to an embodiment of the present disclosure, the memory 130 is a flash memory type, a hard disk type, a multimedia card micro type, or a card type memory (eg, SD or XD memory, etc.), Random Access Memory (RAM), Static Random Access Memory (SRAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Programmable Read (PROM) -Only Memory), a magnetic memory, a magnetic disk, and an optical disk may include at least one type of storage medium. The computing device 100 may operate in relation to a web storage that performs a storage function of the memory 130 on the Internet. The description of the above-described memory is only an example, and the present disclosure is not limited thereto.

The network unit 150 according to an embodiment of the present disclosure may use any type of known wired/wireless communication system.

The network unit 150 may receive a medical image representing a body organ from a medical imaging system. For example, a medical image representing a body organ is a medical image obtained by photographing an upper extremity, and may be data for training or inference of a neural network model. In this case, the medical image in which the body organs are expressed may include all images related to body organs acquired through imaging, such as an X-ray image and a CT image.

In addition, the network unit 150 may transmit and receive information processed by the processor 110 , a user interface, and the like through communication with other terminals. For example, the network unit 150 may provide a user interface generated by the processor 100 to a client (e.g. a user terminal). Also, the network unit 150 may receive an external input of a user authorized as a client and transmit it to the processor 110 . In this case, the processor 100 may process an operation of outputting, modifying, changing, or adding information provided through the user interface based on the user's external input received from the network unit 150 .

Meanwhile, the computing device 100 according to an embodiment of the present disclosure may include a server as a computing system that transmits and receives information through communication with a client. In this case, the client may be any type of terminal that can access the server. For example, the computing device 100 as a server may receive a medical image from a medical image capturing terminal to predict a disease, and may provide a user interface including the predicted result to the user terminal. In this case, the user terminal may output a user interface received from the computing device 100 which is a server, and may receive or process information through interaction with the user.

In an additional embodiment, the computing device 100 may include any type of terminal that receives a data resource generated from an arbitrary server and performs additional information processing.

2 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure;

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network is configured to include at least one or more nodes. Nodes (or neurons) constituting the neural networks may be interconnected by one or more links.

In the neural network, one or more nodes connected through a link may relatively form a relationship between an input node and an output node. The concepts of an input node and an output node are relative, and any node in an output node relationship with respect to one node may be in an input node relationship in a relationship with another node, and vice versa. As described above, an input node-to-output node relationship may be created around a link. One or more output nodes may be connected to one input node through a link, and vice versa.

In the relationship between the input node and the output node connected through one link, the value of the data of the output node may be determined based on data input to the input node. Here, a link interconnecting the input node and the output node may have a weight. The weight may be variable, and may be changed by a user or an algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. An output node value may be determined based on the weight.

As described above, in a neural network, one or more nodes are interconnected through one or more links to form an input node and an output node relationship in the neural network. The characteristics of the neural network may be determined according to the number of nodes and links in the neural network, the correlation between the nodes and the links, and the value of a weight assigned to each of the links. For example, when the same number of nodes and links exist and there are two neural networks having different weight values of the links, the two neural networks may be recognized as different from each other.

A neural network may consist of a set of one or more nodes. A subset of nodes constituting the neural network may constitute a layer. Some of the nodes constituting the neural network may configure one layer based on distances from the initial input node. For example, a set of nodes having a distance of n from the initial input node may constitute n layers. The distance from the initial input node may be defined by the minimum number of links that must be passed to reach the corresponding node from the initial input node. However, the definition of such a layer is arbitrary for description, and the order of the layer in the neural network may be defined in a different way from the above. For example, a layer of nodes may be defined by a distance from the final output node.

The initial input node may mean one or more nodes to which data is directly input without going through a link in a relationship with other nodes among nodes in the neural network. Alternatively, in a relationship between nodes based on a link in a neural network, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the neural network. In addition, the hidden node may mean nodes constituting the neural network other than the first input node and the last output node.

The neural network according to an embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as progresses from the input layer to the hidden layer. can Also, in the neural network according to another embodiment of the present disclosure, the number of nodes in the input layer may be less than the number of nodes in the output layer, and the number of nodes may be reduced as the number of nodes progresses from the input layer to the hidden layer. have. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as the number of nodes progresses from the input layer to the hidden layer. can The neural network according to another embodiment of the present disclosure may be a neural network in a combined form of the aforementioned neural networks.

A deep neural network (DNN) may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer. Deep neural networks can be used to identify the latent structures of data. In other words, it can identify the potential structure of photos, texts, videos, voices, and music (e.g., what objects are in the photos, what the text and emotions are, what the texts and emotions are, etc.) . Deep neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), auto encoders, generative adversarial networks (GANs), and restricted boltzmann machines (RBMs). machine), a deep belief network (DBN), a Q network, a U network, a Siamese network, and a Generative Adversarial Network (GAN). The description of the deep neural network described above is only an example, and the present disclosure is not limited thereto.

In an embodiment of the present disclosure, the network function may include an autoencoder. The auto-encoder may be a kind of artificial neural network for outputting output data similar to input data. The auto encoder may include at least one hidden layer, and an odd number of hidden layers may be disposed between the input/output layers. The number of nodes in each layer may be reduced from the number of nodes of the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically with reduction from the bottleneck layer to the output layer (symmetrically with the input layer). The auto-encoder can perform non-linear dimensionality reduction. The number of input layers and output layers may correspond to a dimension after preprocessing the input data. In the auto-encoder structure, the number of nodes of the hidden layer included in the encoder may have a structure that decreases as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes located between the encoder and decoder) is too small, a sufficient amount of information may not be conveyed, so a certain number or more (e.g., more than half of the input layer, etc.) ) may be maintained.

The neural network may be trained using at least one of supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Learning of the neural network may be a process of applying knowledge for the neural network to perform a specific operation to the neural network.

A neural network can be trained in a way that minimizes output errors. In the training of a neural network, iteratively input the training data into the neural network, calculate the output of the neural network and the target error for the training data, and calculate the error of the neural network from the output layer of the neural network to the input layer in the direction to reduce the error. It is a process of updating the weight of each node in the neural network by backpropagation in the direction. In the case of teacher learning, learning data in which the correct answer is labeled in each learning data is used (ie, labeled learning data), and in the case of comparative learning, the correct answer may not be labeled in each learning data. That is, for example, learning data in the case of teacher learning related to data classification may be data in which categories are labeled in each of the learning data. Labeled training data is input to the neural network, and an error can be calculated by comparing the output (category) of the neural network with the label of the training data. As another example, in the case of comparison learning related to data classification, an error may be calculated by comparing the input training data with the neural network output. The calculated error is back propagated in the reverse direction (ie, from the output layer to the input layer) in the neural network, and the connection weight of each node of each layer of the neural network may be updated according to the back propagation. The change amount of the connection weight of each node to be updated may be determined according to a learning rate. The computation of the neural network on the input data and the backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently according to the number of repetitions of the learning cycle of the neural network. For example, in the early stage of learning of a neural network, a high learning rate can be used to enable the neural network to quickly obtain a certain level of performance, thereby increasing efficiency, and using a low learning rate at a later stage of learning can increase accuracy.

In the training of neural networks, in general, the training data may be a subset of real data (that is, data to be processed using the trained neural network), and thus, the error on the training data is reduced, but the error on the real data is reduced. There may be increasing learning cycles. Overfitting is a phenomenon in which errors on actual data increase by over-learning on training data as described above. For example, a phenomenon in which a neural network that has learned a cat by seeing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing errors in machine learning algorithms. In order to prevent such overfitting, various optimization methods can be used. In order to prevent overfitting, methods such as increasing the training data, regularization, and dropout that deactivate some of the nodes of the network in the process of learning, and the use of a batch normalization layer are applied. can

3 is a block diagram illustrating an operation of a neural network model for reading a state of an object of interest according to an embodiment of the present disclosure.

Referring to FIG. 3 , the reading model 200 according to an embodiment of the present disclosure may generate an output value 20 for reading a state of an interest based on an input image 10 . The input image 10 may include images 11 , 12 , and 13 corresponding to at least a portion of the object of interest, respectively. The output value 20 may include output values 21 , 22 , and 23 corresponding to at least a portion of the object of interest, respectively. For example, the read model 200 may include a first input image 11 including one of the detailed regions of interest, a second input image 12 including the other, and an Nth input image including another one. Each of the input images 13 (N is a natural number) may be input. The reading model 200 may extract features of each of the plurality of input images 11 , 12 , and 13 to generate output values 21 , 22 , and 23 for reading a state of an interest, respectively. The first output value 21 may be basic data for reading the state of the object of interest generated based on the characteristic of the first input image 11 to which the shared characteristic of the input images 11 , 12 , and 13 is reflected. The second output value 22 may be basic data for reading the state of the object of interest generated based on the characteristic of the second input image 12 to which the shared characteristic of the input images 11 , 12 , and 13 is reflected. The Nth output value 23 may be basic data for reading the state of the object of interest generated based on the characteristics of the Nth input image 13 to which the shared characteristics of the input images 11 , 12 , and 13 are reflected.

The read model 200 may include a neural network model that receives a plurality of medical images and outputs a regression value corresponding to each of the plurality of medical images. The reading model 200 may be a model optimized for determining the state of the entire object of interest based on medical images that individually include a part of the object of interest that is distinguished from each other. The reading model 200 may learn and infer the state of the object of interest by reflecting the shared feature based on the correlation between the independent features in the independent features related to the object of interest. In other words, when a full task for reading the state of interest and a sub-task having a correlation for the entire task exist, the reading model 200 uses a plurality of neural networks to generate the entire task based on the sub-task. The task can be performed effectively.

4 is a conceptual diagram illustrating an architecture of a neural network model for reading a state of an object of interest according to an embodiment of the present disclosure. 4 , it is assumed that the reading model according to an embodiment of the present invention is a neural network model for bone age reading.

The reading model according to an embodiment of the present invention may include a first neural network that receives each of a plurality of images including at least a part of the elbow and generates first feature vectors each including local features of the elbow. In this case, the plurality of images include first sub-input images including at least a portion of the elbow photographed in the first direction, and second sub-input images including at least a portion of the elbow photographed in a second direction different from the first direction. It may include sub-input images. For example, the first sub-input images may include frontal images of at least a part of the elbow. At least a part of the elbow included in the first sub-input images may include an olecranon, a trochlea, etc. among detailed bone regions of the elbow. The second sub-input images may include side images of at least a part of the elbow. At least a portion of the elbow included in the second sub-input images may include a lateral condyle, an epicondyle, a proximal radial epiphysis, and the like.

Referring to FIG. 4 , the first neural network receives images 31 , 32 , 33 , 34 in which detailed bone regions are best expressed according to a photographing angle of the elbow, and receives images 31 , 32 , 33 , and 34 that individually represent regional features of the elbow. The feature vectors 41a, 41b, 41c, and 41d may be generated. The first neural network may include sub-neural networks 211 , 212 , 213 , 214 that receive radiographic images 31 , 32 , 33 , 34 including each of the detailed bone regions of the four elbows. can The first sub-neural network A 211 may receive an image 31 including the periphery among detailed bone regions of the elbow and output a first feature vector 41a related to the periphery. The first sub-neural network B 212 receives an image 32 including a lateral condyle and a condyle among detailed bone regions of the elbow, and outputs a first feature vector 41b related to the lateral condyle and condyle. can The first sub-neural network C 213 may receive an image 33 including a trochle among detailed bone regions of the elbow and output a first feature vector 41c related to the trochlear. The first sub-neural network D 214 may receive an image 34 including the proximal radial epiphysis among detailed bone regions of the elbow and output a first feature vector 41d related to the proximal radial epiphysis.

In addition, the read model may include a second neural network that receives a combined feature vector based on the correlation between the first feature vectors and generates a second feature vector including the global feature of the elbow. The read model may identify vectors having a correlation among the first feature vectors output through the first neural network. In this case, the correlation may be determined according to the structure, shape, clinical characteristics, etc. between detailed bone regions of the elbow. The read model may generate the input value of the second neural network by combining the first feature vectors with the correlation. In this case, various methods may be applied to a method for combining the first feature vectors according to characteristics of correlation, such as simple matching and mutually shared feature extraction and merging.

For example, referring to FIG. 4 , the read model includes a first feature vector 41a associated with the periphery, a first feature vector 41b associated with the condyle and the epiglottis, a first feature vector 41c associated with the trochle. The input feature vector 42 of the second neural network 220 may be generated by combining the first feature vector 41d associated with the proximal radial epiphysis. That is, since the four first feature vectors 41a, 41b, 41c, and 41d are correlated in clinical characteristics for judging the bone maturity of the elbow, the read model uses the four first feature vectors 41a, 41d 41b, 41c, and 41d) are combined to generate the input feature vector 42 of the second neural network 220 . If only some of the four first feature vectors 41a, 41b, 41c, and 41d have correlation, the read model may determine whether there is a correlation and combine only some to generate the input feature vector. .

The second neural network 220 receives a feature vector 42 generated by a combination of the four first feature vectors 41a, 41b, 41c, and 41d, and generates a second feature vector 43 representing the global feature of the elbow. can be printed out. In other words, the second neural network 220 generates a second feature vector 43 representing the characteristics of the entire elbow region based on the shared features of the occipital, condyle, condyle, trochle, and proximal radial epiphysis, which are detailed bone regions of the elbow. can create

In addition, the read model may include a third neural network that receives each of the first feature vectors to which the second feature vector is reflected and generates data on the bone maturity of the elbow. The read model may generate an input value of the third neural network by individually combining the second feature vector and the first feature vector determined to be correlated. Each combination of the second feature vector and the first feature vectors may be a process of updating the first feature vector by reflecting features included in the second feature vector to each of the first feature vectors.

For example, referring to FIG. 4 , the read model includes a first feature vector 41a associated with the periphery, a first feature vector 41b associated with the condyle and the epiglottis, a first feature vector 41c associated with the trochle. The input feature vectors 44a, 44b, 44c, and 44d of the third neural network may be generated by combining each of the first feature vectors 41d associated with the proximal radial epiphysis and the second feature vector 43 . The third neural network receives the feature vectors 44a, 44b, 44c, and 44d of the detailed bone regions in which the global characteristic of the elbow is reflected, and scores 51, 52, 53, 54) can be created. The third neural network may include sub neural networks 231 , 232 , 233 , and 234 that receive input feature vectors 44a , 44b , 44c , and 44d corresponding to each of the detailed bone regions of the four elbows. The third sub-neural network A 231 may receive the periphery-related feature vector 44a to which the second feature vector 43 is reflected, and may output a first score 51 indicating the bone maturity of the periphery. The third sub-neural network B 232 receives the lateral condyle and the epicondyle-related feature vector 44b to which the second feature vector 43 is reflected, and receives a second score 52 indicating the degree of bone maturity of the lateral condyle and the epicondyle. can be printed out. The third sub-neural network C 233 may receive the trochlear related feature vector 44c to which the second feature vector 43 is reflected and output a third score 53 indicating the bone maturity of the trochle. The third sub-neural network D 234 may receive the proximal radial epiphysis-related feature vector 44d to which the second feature vector 43 is reflected, and may output a fourth score 54 indicating the degree of bone maturity of the proximal radial epiphysis. The scores 51 , 52 , 53 , and 54 for each detailed bone region of the elbow output through the third neural network may be used to read the bone age of the elbow based on the bone maturity.

5 is a block diagram illustrating a process of reading a state of an object of interest by a computing device according to an embodiment of the present disclosure.

Referring to FIG. 5 , the processor 110 of the computing device 100 according to an embodiment of the present disclosure uses the pre-processing model 300 to generate an input image 62 of the read model 200 , Processing may be performed on the included medical image 61 . The medical image 61 may be a radiographic image captured to include an object of interest corresponding to a body composition such as an upper limb, and may include all images captured by various techniques in addition to this. In this case, the pre-processing model 300 is a first pre-processing module 310 for normalization of the medical image 61 , segmenting the medical image so that at least a portion of the object of interest is included in the medical image 61 . A second pre-processing module 320 may be included. Each of the first pre-processing module 310 and the second pre-processing module 320 may be a neural network model optimized for image processing.

For example, the processor 110 may detect a plurality of reference points in the region of interest existing in the medical image 61 using the first pre-processing module 310 . The processor 110 may perform geometric transformation on the object of interest existing in the medical image 61 based on the plurality of reference points by using the first pre-processing module 310 . The normalization process by the first pre-processing module 310 can be understood as a process of unifying the expression of the object of interest that exists in various forms according to the shooting angle, direction, etc. into a form suitable for input of the reading model 200 .

In addition, the processor 110 may detect a center point of a bounding box including at least a portion of the object of interest based on the normalized medical image using the second preprocessing module 320 . The processor 110 may determine the center point of the bounding box as a keypoint for generating the input image 62 . The processor 110 may generate the input image 62 by segmenting a predetermined region centered on the key point from the normalized medical image using the second preprocessing module 320 . In this case, the predetermined area may be predetermined to have a size suitable for input of the read model 200 . The processing process by the second pre-processing module 320 as described above is a process of cropping a plurality of images including each of the detailed regions of interest corresponding to the input image 62 from the normalized medical image and generating it. can be understood If the medical image 61 is already a normalized image, the processor 110 generates the input image 62 from the medical image 61 using the second pre-processing module 320 without the operation of the first pre-processing module 310 . You can also create

The processor 110 may generate the output data 63 in the form of a score based on the input image 62 including a plurality of images including at least a portion of the ROI by using the reading model 200 . The processor 110 may use the reading model 200 to extract features regarding at least a portion of the object of interest and all of the object of interest. The processor 110 may generate the basic data 63 necessary for determining the state of the interest in the form of a score based on the above-described extracted features using the reading model 200 . Since the details of feature extraction and score calculation of the read model 200 have been described in detail with reference to FIGS. 3 and 4 , further descriptions will be omitted.

The processor 110 may generate the state reading result 65 of the interest by reflecting the data 63 in the form of a score, which is an output value of the reading model 200 , to the data 64 related to a biological indicator such as gender. For example, if it is assumed that the state of interest to be read by the processor 110 is the bone age of the elbow, the processor 110 uses the reading model 200 based on images including detailed bone regions of the elbow, respectively. Scores representing detailed bone regions of each detailed bone region may be calculated. The processor 110 may sum the scores of each detailed bone region calculated through the read model in order to determine the bone age of the entire elbow. The processor 110 may estimate the final bone age of the elbow itself by reflecting the summation result in a graph or function indicating a change according to gender.

6 is a conceptual diagram illustrating an operation and configuration of a preprocessing model of a computing device according to an embodiment of the present disclosure. 6 , it is assumed that the preprocessing model according to an embodiment of the present invention is a neural network model for converting an elbow image.

Referring to FIG. 6 , the first pre-processing module 310 according to an embodiment of the present invention receives a medical image 71 in which the upper extremities including the elbow are captured, and a front image 72 of the elbow or a side image of the elbow ( 73), the medical image 71 may be converted. Since the medical image 71 can be taken from various angles and directions as well as from the front or side, the first pre-processing module 310 performs geometric deformation of the elbow region in order to create images suitable for input of the elbow bone age reading model. Through this, the medical image 71 may be normalized. The first pre-processing module 310 detects three reference points in the medical image 71 and calculates parameters of an image transformation function using the three reference points to convert the medical image 71 into the front image 72 or the side image 73 . ) can be converted to In this case, an affine transform may be applied to the image transformation method of the first pre-processing module 310 , but the present disclosure is not limited thereto.

Meanwhile, the first preprocessing module 310 may receive images labeled with reference points for calculating an image transformation function and train the neural network. The first preprocessing module 310 may convert each of the labeled reference points into a probability map following a Gaussian distribution centered on the training image. The first preprocessing module 310 may train the neural network to calculate a probability map corresponding to each of the reference points. As a loss function for learning, a binary cross-entropy loss function may be used. The above-described learning method is only an example, and the present disclosure is not limited thereto.

The second pre-processing module 320 according to an embodiment of the present invention receives the front image 72 and the side image 73 of the elbow generated through the first pre-processing module 310 for each detailed bone region of the elbow. Images 74 , 75 , 76 , and 77 may be generated. The second pre-processing module 320 segments the front image 72 and the side image 73 of the elbow, respectively, to obtain images 74, 75, 76, and 77 including four elbow regions limited to a predetermined area. can create The second pre-processing module 320 detects the center point of the peritoneal region and the trochlear region in the frontal image 72 of the elbow, and crops the frontal image 72 based on the central point to obtain the peritoneal image 74 and the trochlear image 75 ) can be created. The second pre-processing module 320 detects the center point of the lateral condyle, the supracondylar region, and the proximal radial epiphysis in the lateral image 73 of the elbow, and crops the front image 72 based on the central point to thereby crop the lateral condyle. And an epiphyseal image 76 and a proximal radial epiphyseal image 77 may be generated. In this case, the center point in the front image 72 and the side image 73 may be detected based on a bounding box set based on the learning image. The images 74 , 75 , 76 , and 77 of the four detailed bone regions of the elbow generated by the second preprocessing module 320 may be used as input to a model for reading the age of the elbow bone.

Meanwhile, the second preprocessing module 320 may receive segmentation images for each bone part of the elbow to learn the neural network. The second pre-processing module 320 may calculate a bounding box for each of the four detailed bone regions (eg, periphery, trochlear, etc.) of each of the segmented images. The second preprocessing module 320 may train the neural network by using the calculated center point of the bounding box as a keypoint label. As a loss function for learning, a binary cross-entropy loss function may be used. The above-described learning method is only an example, and the present disclosure is not limited thereto.

7 is a flowchart of a method for reading a state of an object of interest based on a medical image according to an embodiment of the present disclosure.

Referring to FIG. 7 , in step S100 , the computing device 100 may receive medical images of an object of interest from a medical imaging system. For example, the medical image may be an X-ray image obtained by photographing the upper extremities centered on the elbow region. The computing device 100 may generate a plurality of model input images including at least a portion of the object of interest based on the medical images. For example, the computing device 100 may convert an elbow of interest to have a front view or a side view through geometric deformation of medical images captured by the upper limb. The computing device 100 includes a plurality of images including each of the detailed bone regions of the elbow (the periphery, trochle, lateral condyle, and proximal radial epiphysis) from the normalized medical images in which the view of the elbow of interest is transformed. can create

In operation S200 , the computing device 100 may extract first feature vectors indicating the local feature of the interest from each of the images generated in operation S100 using the read model. For example, the computing device 100 may extract features of each of the four detailed bone regions of the elbow using the read model. Specifically, the read model may receive images of the periphery, trochlear, lateral condyle, and proximal radial epiphysis to generate first feature vectors representing regional characteristics of the elbow bone region.

In operation S300 , the computing device 100 may generate a second feature vector representing the global feature of the interest based on the first feature vectors generated in operation S200 using the read model. The computing device 100 may identify vectors having a correlation among the first feature vectors using the read model, and generate a second feature vector based on a result of combining them. For example, the computing device 100 may combine all of the first feature vectors corresponding to the four types of correlated detailed bone regions of the elbow, such as the periphery, trochlear, lateral condyle, and epicondyle, and the proximal radius of the epiphysis. . The computing device 100 uses the read model to generate a second feature vector that is a global feature vector of the elbow bone region including the shared feature of each detailed bone region based on the first feature vectors including the feature of each detailed bone region can create

In step S400, the computing device 100 combines the second feature vector generated in step S300 and the first feature vectors generated in step S200 to obtain regional feature vectors (i.e. second) in which the global feature of the object of interest is reflected. first feature vectors to which the feature vector is reflected) may be generated. The computing device 100 may use the read model to generate state read data for at least a portion of the object of interest based on each of the first feature vectors to which the second feature vector is reflected. The computing device 100 may finally determine the overall state of the object of interest based on the state reading data and the biometric index data of at least a portion of the object of interest. For example, the computing device 100 may use the read model to indicate the bone maturity for each of the occipital, trochlear, lateral condylar and epicondylar regions, and the proximal radial epiphysis based on the first feature vectors to which the second feature vector is reflected. You can create a score. Here, the degree of bone maturity may be predetermined based on an index indicating a grade of bone maturity for each region. The computing device 100 may generate a score indicating the bone maturity of the entire elbow bone region by summing the scores indicating the bone maturity for each sub-region. The computing device 100 may calculate the bone age of the entire elbow bone region by reflecting the score indicating the bone maturity of the entire elbow bone region to a gender-based bone age conversion function.

8 is a flowchart of a pre-processing method of a medical image according to an embodiment of the present disclosure.

Referring to FIG. 8 , in step S110 , the computing device 100 may detect a plurality of reference points for geometric deformation of the object of interest based on the region in which the object of interest exists in at least one medical image using the preprocessing model. have. For example, the computing device 100 may detect three reference points for the elbow bone region of interest in a medical image in which the upper limb including the elbow bone region is captured by using the preprocessing model. The three reference points can be used for geometrical deformation of the medical image to align the elbow bone region in a front view or a side view as a basis.

In operation S120 , the computing device 100 may normalize at least one medical image based on the plurality of reference points detected in operation S110 using the preprocessing model. The computing device 100 may perform geometric transformation of a medical image by calculating a parameter of an image transformation function based on a plurality of reference points using a preprocessing model. For example, the computing device 100 may transform the medical image into a front image or a side image of the elbow bone region based on three reference points of the elbow bone region.

In operation S130 , the computing device 100 may detect the center point of the bounding box including at least a portion of the object of interest based on the medical images normalized in operation S120 using the preprocessing model. The computing device 100 may set a bounding box for each of the detailed regions constituting the object of interest by using the preprocessing model. The computing device 100 may detect the center points of the bounding box as a reference for segmentation of the medical image by using the preprocessing model. For example, the computing device 100 may detect center points of bounding boxes including each of the four detailed elbow bone regions from normalized medical images. In this case, the center point of the bounding box including the periphery and the center point of the bounding box including the trochlear may be detected from the frontal images of the elbow among the normalized medical images. Among the normalized medical images, the center point of the bounding box including the lateral condyle and the condyle and the center point of the bounding box including the proximal radial epiphysis may be detected from the side images of the elbow.

In operation S140 , the computing device 100 may generate a plurality of input images including a predetermined region from the normalized medical images based on the center points detected in operation S130 using the preprocessing model. The computing device 100 may segment normalized medical images by determining predetermined regions determined based on the center points of the bounding box as input images using the preprocessing model. The computing device 100 may use images generated by segmentation from normalized medical images to read the state of the object of interest. For example, the computing device 100 may generate an input image of the peritoneal region and an input image of the trochlear region determined based on center points of the bounding boxes from the front images of the elbow. The computing device 100 may generate an input image of a lateral condyle and an epicondylar region and an input image of a proximal radius epiphysis region determined based on center points of the bounding boxes from the lateral images of the elbow. In this case, the size of each input image may be determined according to a predetermined range area based on the central point.

On the other hand, a computer-readable medium storing a data structure according to an embodiment of the present disclosure is disclosed.

The data structure may refer to the organization, management, and storage of data that enables efficient access and modification of data. A data structure may refer to an organization of data to solve a specific problem (eg, data retrieval, data storage, and data modification in the shortest time). A data structure may be defined as a physical or logical relationship between data elements designed to support a particular data processing function. The logical relationship between data elements may include a connection relationship between user-defined data elements. Physical relationships between data elements may include actual relationships between data elements physically stored on a computer-readable storage medium (eg, persistent storage). A data structure may specifically include a set of data, relationships between data, and functions or instructions applicable to data. Through an effectively designed data structure, the computing device can perform an operation while using the resource of the computing device to a minimum. Specifically, the computing device may increase the efficiency of operations, reads, insertions, deletions, comparisons, exchanges, and retrievals through effectively designed data structures.

A data structure may be classified into a linear data structure and a non-linear data structure according to the type of the data structure. The linear data structure may be a structure in which only one piece of data is connected after one piece of data. The linear data structure may include a list, a stack, a queue, and a deck. A list may mean a set of data in which an order exists internally. The list may include a linked list. The linked list may be a data structure in which data is linked in such a way that each data is linked in a line with a pointer. In a linked list, a pointer may contain information about a link with the next or previous data. A linked list may be expressed as a single linked list, a doubly linked list, or a circularly linked list according to a shape. A stack can be a data enumeration structure with limited access to data. A stack can be a linear data structure in which data can be manipulated (eg, inserted or deleted) at only one end of the data structure. The data stored in the stack may be a data structure (LIFO-Last in First Out). A queue is a data listing structure that allows limited access to data. Unlike a stack, the queue may be a data structure that comes out later (FIFO-First in First Out) as data stored later. A deck can be a data structure that can process data at either end of the data structure.

The non-linear data structure may be a structure in which a plurality of data is connected after one data. The nonlinear data structure may include a graph data structure. A graph data structure may be defined as a vertex and an edge, and the edge may include a line connecting two different vertices. A graph data structure may include a tree data structure. The tree data structure may be a data structure in which one path connects two different vertices among a plurality of vertices included in the tree. That is, it may be a data structure that does not form a loop in the graph data structure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. Hereinafter, the neural network is unified and described. The data structure may include a neural network. And the data structure including the neural network may be stored in a computer-readable medium. Data structures, including neural networks, also include preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, and neural networks. It may include a loss function for learning of . A data structure comprising a neural network may include any of the components disclosed above. That is, the data structure including the neural network includes preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation function associated with each node or layer of the neural network, and neural network It may be configured to include all or any combination thereof, such as a loss function for learning of . In addition to the above-described configurations, a data structure including a neural network may include any other information that determines a characteristic of a neural network. In addition, the data structure may include all types of data used or generated in the operation process of the neural network, and is not limited to the above. Computer-readable media may include computer-readable recording media and/or computer-readable transmission media. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network is configured to include at least one or more nodes.

The data structure may include data input to the neural network. A data structure including data input to the neural network may be stored in a computer-readable medium. The data input to the neural network may include learning data input in a neural network learning process and/or input data input to the neural network in which learning is completed. Data input to the neural network may include pre-processing data and/or pre-processing target data. The preprocessing may include a data processing process for inputting data into the neural network. Accordingly, the data structure may include data to be pre-processed and data generated by pre-processing. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

The data structure may include the weights of the neural network. (In this specification, weight and parameter may be used interchangeably.) And the data structure including the weight of the neural network may be stored in a computer-readable medium. The neural network may include a plurality of weights. The weight may be variable, and may be changed by a user or an algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. A data value output from the output node may be determined based on the weight. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

By way of example and not limitation, the weight may include a weight variable in a neural network learning process and/or a weight in which neural network learning is completed. The variable weight in the neural network learning process may include a weight at a time point at which a learning cycle starts and/or a weight variable during the learning cycle. The weight for which neural network learning is completed may include a weight for which a learning cycle is completed. Accordingly, the data structure including the weights of the neural network may include a data structure including the weights that vary in the neural network learning process and/or the weights on which the neural network learning is completed. Therefore, it is assumed that the above-described weights and/or combinations of weights are included in the data structure including the weights of the neural network. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

The data structure including the weights of the neural network may be stored in a computer-readable storage medium (eg, memory, hard disk) after being serialized. Serialization can be the process of converting a data structure into a form that can be reconstructed and used later by storing it on the same or a different computing device. The computing device may serialize the data structure to send and receive data over the network. A data structure including weights of the serialized neural network may be reconstructed in the same computing device or in another computing device through deserialization. The data structure including the weight of the neural network is not limited to serialization. Furthermore, the data structure including the weights of the neural network is a data structure to increase computational efficiency while using the resources of the computing device to a minimum (e.g., B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree). The foregoing is merely an example, and the present disclosure is not limited thereto.

The data structure may include hyper-parameters of the neural network. In addition, the data structure including the hyperparameters of the neural network may be stored in a computer-readable medium. The hyper parameter may be a variable variable by a user. Hyperparameters are, for example, learning rate, cost function, number of iterations of the learning cycle, weight initialization (e.g., setting the range of weight values to be initialized for weights), Hidden Unit The number (eg, the number of hidden layers, the number of nodes of the hidden layer) may be included. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

9 is a simplified, general schematic diagram of an example computing environment in which embodiments of the present disclosure may be implemented.

Although the present disclosure has been described above as being generally capable of being implemented by a computing device, those skilled in the art will appreciate that the present disclosure may be implemented in hardware and software in combination with computer-executable instructions and/or other program modules and/or in combination with computer-executable instructions and/or other program modules that may be executed on one or more computers. It will be appreciated that it can be implemented as a combination.

Generally, program modules include routines, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types. In addition, those skilled in the art will appreciate that the methods of the present disclosure can be applied to single-processor or multiprocessor computer systems, minicomputers, mainframe computers as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, and the like. It will be appreciated that each of these may be implemented in other computer system configurations, including those capable of operating in connection with one or more associated devices.

The described embodiments of the present disclosure may also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Computers typically include a variety of computer-readable media. Any medium accessible by a computer can be a computer readable medium, and such computer readable media includes volatile and nonvolatile media, transitory and non-transitory media, removable and non-transitory media. including removable media. By way of example, and not limitation, computer-readable media may include computer-readable storage media and computer-readable transmission media. Computer readable storage media includes volatile and nonvolatile media, temporary and non-transitory media, 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. includes media. A computer-readable storage medium may be RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage device, magnetic cassette, magnetic tape, magnetic disk storage device, or other magnetic storage device. device, or any other medium that can be accessed by a computer and used to store the desired information.

Computer readable transmission media typically embodies computer readable instructions, data structures, program modules or other data, etc. in a modulated data signal such as a carrier wave or other transport mechanism, and Includes any information delivery medium. The term modulated data signal means a signal in which one or more of the characteristics of the signal is set or changed so as to encode information in the signal. By way of example, and not limitation, computer-readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.

An example environment 1100 implementing various aspects of the disclosure is shown including a computer 1102 , the computer 1102 including a processing unit 1104 , a system memory 1106 , and a system bus 1108 . do. A system bus 1108 couples system components, including but not limited to system memory 1106 , to the processing device 1104 . The processing device 1104 may be any of a variety of commercially available processors. Dual processor and other multiprocessor architectures may also be used as processing unit 1104 .

The system bus 1108 may be any of several types of bus structures that may further interconnect a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures. System memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) 1112 . A basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, EEPROM, etc., the BIOS is the basic input/output system (BIOS) that helps transfer information between components within computer 1102, such as during startup. contains routines. RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.

The computer 1102 may also include an internal hard disk drive (HDD) 1114 (eg, EIDE, SATA) - this internal hard disk drive 1114 may also be configured for external use within a suitable chassis (not shown). Yes—a magnetic floppy disk drive (FDD) 1116 (eg, for reading from or writing to removable diskette 1118), and an optical disk drive 1120 (eg, a CD-ROM) for reading from, or writing to, disk 1122, or other high capacity optical media, such as DVD. The hard disk drive 1114 , the magnetic disk drive 1116 , and the optical disk drive 1120 are connected to the system bus 1108 by the hard disk drive interface 1124 , the magnetic disk drive interface 1126 , and the optical drive interface 1128 , respectively. ) can be connected to The interface 1124 for implementing an external drive includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

These drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like. In the case of computer 1102, drives and media correspond to storing any data in a suitable digital format. Although the description of computer readable media above refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will use zip drives, magnetic cassettes, flash memory cards, cartridges, etc. It will be appreciated that other tangible computer-readable media such as etc. may also be used in the exemplary operating environment and any such media may include computer-executable instructions for performing the methods of the present disclosure.

A number of program modules may be stored in the drive and RAM 1112 , including an operating system 1130 , one or more application programs 1132 , other program modules 1134 , and program data 1136 . All or portions of the operating system, applications, modules, and/or data may also be cached in RAM 1112 . It will be appreciated that the present disclosure may be implemented in various commercially available operating systems or combinations of operating systems.

A user may enter commands and information into the computer 1102 via one or more wired/wireless input devices, for example, a pointing device such as a keyboard 1138 and a mouse 1140 . Other input devices (not shown) may include a microphone, IR remote control, joystick, game pad, stylus pen, touch screen, and the like. Although these and other input devices are connected to the processing unit 1104 through an input device interface 1142 that is often connected to the system bus 1108, parallel ports, IEEE 1394 serial ports, game ports, USB ports, IR interfaces, It may be connected by other interfaces, etc.

A monitor 1144 or other type of display device is also coupled to the system bus 1108 via an interface, such as a video adapter 1146 . In addition to the monitor 1144, the computer typically includes other peripheral output devices (not shown), such as speakers, printers, and the like.

Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 1148 , via wired and/or wireless communications. Remote computer(s) 1148 may be workstations, computing device computers, routers, personal computers, portable computers, microprocessor-based entertainment devices, peer devices, or other common network nodes, and are typically connected to computer 1102 . Although it includes many or all of the components described for it, only memory storage device 1150 is shown for simplicity. The logical connections shown include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, eg, a wide area network (WAN) 1154 . Such LAN and WAN networking environments are common in offices and companies, and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, for example, the Internet.

When used in a LAN networking environment, the computer 1102 is coupled to the local network 1152 through a wired and/or wireless communication network interface or adapter 1156 . Adapter 1156 may facilitate wired or wireless communication to LAN 1152 , which LAN 1152 also includes a wireless access point installed therein for communicating with wireless adapter 1156 . When used in a WAN networking environment, the computer 1102 may include a modem 1158, be connected to a communication computing device on the WAN 1154, or establish communications over the WAN 1154, such as over the Internet. have other means. A modem 1158 , which may be internal or external and a wired or wireless device, is coupled to the system bus 1108 via a serial port interface 1142 . In a networked environment, program modules described for computer 1102 , or portions thereof, may be stored in remote memory/storage device 1150 . It will be appreciated that the network connections shown are exemplary and other means of establishing a communication link between the computers may be used.

The computer 1102 may be associated with any wireless device or object that is deployed and operates in wireless communication, for example, a printer, scanner, desktop and/or portable computer, portable data assistant (PDA), communication satellite, wireless detectable tag. It operates to communicate with any device or place, and phone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure as in a conventional network or may simply be an ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) makes it possible to connect to the Internet, etc. without a wired connection. Wi-Fi is a wireless technology such as cell phones that allows these devices, eg, computers, to transmit and receive data indoors and outdoors, ie anywhere within range of a base station. Wi-Fi networks use a radio technology called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks may operate in unlicensed 2.4 and 5 GHz radio bands, for example at 11 Mbps (802.11a) or 54 Mbps (802.11b) data rates, or in products that include both bands (dual band). .

One of ordinary skill in the art of this disclosure will understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, instructions, information, signals, bits, symbols, and chips that may be referenced in the above description are voltages, currents, electromagnetic waves, magnetic fields or particles, optical field particles or particles, or any combination thereof.

Those of ordinary skill in the art of the present disclosure will recognize that the various illustrative logical blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein include electronic hardware, (convenience For this purpose, it will be understood that it may be implemented by various forms of program or design code (referred to herein as software) or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. A person skilled in the art of the present disclosure may implement the described functionality in various ways for each specific application, but such implementation decisions should not be interpreted as a departure from the scope of the present disclosure.

The various embodiments presented herein may be implemented as methods, apparatus, or articles of manufacture using standard programming and/or engineering techniques. The term article of manufacture includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, computer-readable storage media include magnetic storage devices (eg, hard disks, floppy disks, magnetic strips, etc.), optical disks (eg, CDs, DVDs, etc.), smart cards, and flash drives. memory devices (eg, EEPROMs, cards, sticks, key drives, etc.). Also, various storage media presented herein include one or more devices and/or other machine-readable media for storing information.

It is to be understood that the specific order or hierarchy of steps in the presented processes is an example of exemplary approaches. Based on design priorities, it is to be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present disclosure. The appended method claims present elements of the various steps in a sample order, but are not meant to be limited to the specific order or hierarchy presented.

The description of the presented embodiments is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure is not intended to be limited to the embodiments presented herein, but is to be construed in the widest scope consistent with the principles and novel features presented herein.

Claims (12)

A method of reading a status of a subject of interest based on a medical image performed by a computing device including at least one processor, the method comprising:
extracting first feature vectors from each of a plurality of input images including at least a portion of an object of interest by using the pre-trained reading model;
generating a second feature vector based on a correlation between first feature vectors corresponding to each of the plurality of input images by using the read model; and
using the read model, combining each of the first feature vectors with the second feature vector, and determining a state of the interest based on the combined feature vectors;
containing,
Way.
The method of claim 1,
The first feature vectors are
each comprising a regional feature vector for the object of interest generated based on at least a portion of the object of interest, and
The second feature vector is,
a global feature vector for the object of interest generated based on a combination of at least some of the first feature vectors;
Way.
The method of claim 1,
The plurality of input images are
first sub-input images including at least a portion of an object of interest photographed based on a first direction; and
second sub-input images including at least a portion of the object of interest photographed based on a second direction different from the first direction;
containing,
Way.
4. The method of claim 3,
The first sub-input images are
including frontal images of at least a portion of the body region corresponding to the object of interest, and
The second sub-input images are
comprising side images of at least a portion of the body region,
Way.
The method of claim 1,
The step of generating the second feature vector comprises:
determining a correlation between the first feature vectors using the read model, and selecting first feature vectors having a correlation based on a result of the judgment; and
generating a second feature vector based on the first feature vectors in which the selected correlation exists, using the read model;
containing,
Way.
The method of claim 1,
Determining the state of the object of interest comprises:
deriving, using the read model, scores for determining the state of interest based on each of the first feature vectors combined with the second feature vector; and
summing up scores corresponding to each of the first feature vectors, and reflecting a result of the summation in biomarker data related to the subject of interest to read a state of the object of interest;
containing,
Way.
7. The method of claim 6,
The scores for determining the state of the interest are,
scores indicative of bone maturity for at least a portion of the body region corresponding to the subject of interest, and
The state of interest is,
which is the bone age of the body region corresponding to the object of interest,
Way.
The method of claim 1,
The read model is
a first neural network that receives each of the plurality of input images and generates first feature vectors related to at least a part of the object of interest;
a second neural network that receives a combined feature vector based on the correlation between the first feature vectors and generates a second feature vector related to the object of interest; and
a third neural network that receives each of the first feature vectors to which the second feature vector is reflected and generates data on the state of the interest;
containing,
Way.
The method of claim 1,
detecting a center point of a bounding box including at least a portion of the object of interest based on at least one medical image including the object of interest using a pre-trained pre-processing model; and
generating a plurality of input images including a predetermined region determined based on the central point from the at least one medical image by using the preprocessing model;
further comprising,
Way.
10. The method of claim 9,
detecting a plurality of reference points for geometric deformation of the object of interest based on a region in which the object of interest exists in the at least one medical image using the preprocessing model; and
normalizing the at least one medical image based on the plurality of reference points using the preprocessing model;
further comprising,
Way.
A computer program stored in a computer readable storage medium, wherein the computer program, when executed on one or more processors, performs the following operations for reading a condition of a subject of interest based on a medical image, the operations comprising:
extracting first feature vectors from each of a plurality of input images including at least a portion of an object of interest by using the pre-trained reading model;
generating a second feature vector based on a correlation between first feature vectors corresponding to each of the plurality of input images by using the read model; and
combining each of the first feature vectors with the second feature vector using the read model, and determining the state of the interest based on the combined feature vectors;
containing,
A computer program stored in a computer-readable storage medium.
A computing device for reading a condition of a subject of interest based on a medical image, comprising:
a processor including at least one core;
a memory containing program codes executable by the processor; and
a network unit for receiving a medical image;
including,
The processor is
extracting first feature vectors from each of a plurality of input images including at least a portion of an object of interest using the pre-trained reading model;
generating a second feature vector based on a correlation between first feature vectors corresponding to each of the plurality of input images by using the read model;
combining each of the first feature vectors with the second feature vector using the read model, and determining the state of the interest based on the combined feature vectors;
Device.

Images