Electronic device to which machine learning ensemble model for malaria diagnosis and detection is applicable, and operation method thereof

Abstract

An electronic device to which a machine learning ensemble model is applicable, according to an exemplary embodiment of the present disclosure, may comprise: a first model neural-network-trained on the basis of training images; a second model trained on the basis of training test data; a third model using voting based on a plurality of statistical methodologies so as to output the infection probability of malaria infection; and a processor, which inputs, into the first model, a blood image to be tested, uses the first model to make a primary determination of malaria infection about the blood image, inputs, into the second model, blood data to be tested, uses the second model to make a secondary determination of malaria infection about the blood data, and uses the third model to output the malaria infection probability.

Description

Electronic device and its operating method applicable to machine learning ensemble model for malaria diagnosis and detection

[0001] The present disclosure relates to an electronic device and an operating method thereof to which a machine learning ensemble model for malaria diagnosis and detection can be applied.

[0002] Tertiary malaria, caused by the parasite Plasmodium vivax, is a frequent epidemic in Korea. Malaria infection can cause complications such as splenic rupture due to enlargement of the spleen, and central nervous system abnormalities such as memory loss, convulsions, and schizophrenia. Therefore, technologies for rapid and accurate malaria detection are essential.

[0003] Although technologies such as Patent Application No. 10-2018-0071448 have been proposed for the diagnosis and detection of tertian malaria, most of them rely on traditional medical detection methods such as diagnostic reagents and kits.

[0004] The purpose of this invention is to provide a device and method for an ensemble model that combines a convolutional neural network (CNN) model based on object detection based on peripheral blood slide photos of malaria patients and a machine learning model based on blood test results and clinical information for malaria diagnosis, for accurate diagnosis and rapid detection of malaria diseases such as tertian malaria.

[0005] An electronic device to which a machine learning ensemble model according to the present disclosure is applicable for solving the above-described problem may include: a first model trained on a neural network based on a training image; a second model trained on the basis of training test data; a third model configured to output an infection probability regarding malaria infection by voting on a plurality of statistical methodologies; and a processor configured to input a blood image as a test subject into the first model, perform a primary determination on malaria infection for the blood image using the first model, input blood data as a test subject into the second model, perform a secondary determination on malaria infection for the blood data using the second model, and output the malaria infection probability using the third model.

[0006] In addition, a malaria diagnosis and detection method based on a machine learning ensemble model performed by an electronic device according to the present disclosure for solving the above-described problem may include the steps of: inputting a blood image as a test subject into a first model trained by a neural network based on a training image; performing a primary determination as to whether the blood image is infected with malaria using the trained first model; inputting blood data as a test subject into a second model trained based on training test data; performing a secondary determination as to whether the blood data is infected with malaria using the trained second model; and outputting a malaria infection probability using a third model that applies a voting method to a plurality of statistical methodologies.

[0007] In addition, other methods for implementing the present invention, other systems, and computer-readable recording media recording a computer program for executing the above methods may be further provided.

[0008] In addition, a computer program may be further provided that is stored in a medium so that a method for implementing the present invention is performed on a computer.

[0009] According to an electronic device and its operating method applicable to a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure, rapid diagnosis and accurate detection can be achieved when malaria infection is suspected.

[0010] In addition, according to an electronic device and an operating method thereof to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure is applicable, the relative number of malaria parasites according to the number of red blood cells can be counted through deep learning-based real-time object recognition technology, thereby enabling accurate reporting of the degree of malaria infection, and in order to supplement the detection of malaria parasites that deep learning cannot accurately detect, more accurate malaria parasite detection can be achieved by utilizing machine learning based on blood test results.

[0011] The effects of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

[0012] FIG. 1 is a block diagram illustrating an electronic device to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied.

[0013] FIG. 2 is a block diagram illustrating in detail an artificial neural network applied to a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure.

[0014] FIG. 3 is a diagram illustrating the performance of an electronic device to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied.

[0015] FIG. 4 is a diagram illustrating the performance of an electronic device to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied.

[0016] FIG. 5 is a diagram showing the performance of an electronic device to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied.

[0017] FIGS. 6A to 6C are diagrams illustrating type-specific object recognition using a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure.

[0018] FIG. 7 is a block diagram detailing an artificial neural network applied to a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure.

[0019] FIG. 8 is a block diagram illustrating an electronic device to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied.

[0020] Throughout this disclosure, the same reference numerals denote the same components. This disclosure does not describe all elements of the embodiments, and any content that is common in the technical field to which this disclosure belongs or that overlaps between embodiments is omitted. The terms "part, module, element, block" used in the specification may be implemented in software or hardware, and depending on the embodiments, multiple "parts, modules, elements, blocks" may be implemented as a single component, or a single "part, module, element, block" may include multiple components. Throughout the specification, when a part is said to be "connected" to another part, this includes not only cases where it is directly connected, but also cases where it is indirectly connected, and an indirect connection includes a connection via a wireless communication network.

[0021] Additionally, when a part is said to "include" a component, this does not mean that it excludes other components, but rather that it may include other components, unless otherwise specifically stated.

[0022] Throughout the specification, when we say that an element is "on" another element, this includes not only cases where the element is in contact with the other element, but also cases where another element exists between the two elements.

[0023] The terms first, second, etc. are used to distinguish one component from another, and the components are not limited by the aforementioned terms.

[0024] Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0025] The identification codes for each step are used for convenience of explanation and do not describe the order of each step. Each step may be performed in a different order than specified unless the context clearly indicates a specific order.

[0026] The operating principle and embodiments of the present disclosure are described below with reference to the attached drawings.

[0027] As used herein, the term "device according to the present disclosure" encompasses a variety of devices capable of performing computational processing and providing results to a user. For example, the device according to the present disclosure may include a computer, a server device, and a portable terminal, or may be any one of them.

[0028] Here, the computer may include, for example, a notebook, desktop, laptop, tablet PC, slate PC, etc. equipped with a web browser.

[0029] The above server device is a server that processes information by communicating with an external device, and may include an application server, a computing server, a database server, a file server, a game server, a mail server, a proxy server, and a web server.

[0030] The above portable terminal may include, for example, a wireless communication device that ensures portability and mobility, and may include all kinds of handheld-based wireless communication devices such as a PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), WiBro (Wireless Broadband Internet) terminal, a smart phone, and a wearable device such as a watch, a ring, a bracelet, an anklet, a necklace, glasses, contact lenses, or a head-mounted device (HMD).

[0031] The artificial intelligence-related functions according to the present disclosure are operated via a processor and memory. The processor may be comprised of one or more processors. In this case, one or more processors may be a general-purpose processor such as a CPU, an AP, a Digital Signal Processor (DSP), a graphics-only processor such as a GPU or a Vision Processing Unit (VPU), or an artificial intelligence-only processor such as an NPU. One or more processors control the processing of input data according to predefined operating rules or artificial intelligence models stored in memory. Alternatively, if one or more processors are artificial intelligence-only processors, the artificial intelligence-only processor may be designed with a hardware structure specialized for processing a specific artificial intelligence model.

[0032] The artificial intelligence-related functions according to the present disclosure are operated via a processor and memory. The processor may be comprised of one or more processors. In this case, one or more processors may be a general-purpose processor such as a CPU, an AP, a Digital Signal Processor (DSP), a graphics-only processor such as a GPU or a Vision Processing Unit (VPU), or an artificial intelligence-only processor such as an NPU. One or more processors control the processing of input data according to predefined operating rules or artificial intelligence models stored in memory. Alternatively, if one or more processors are artificial intelligence-only processors, the artificial intelligence-only processor may be designed with a hardware structure specialized for processing a specific artificial intelligence model.

[0033] An artificial intelligence model is characterized by being created through learning. Here, being created through learning means that a basic artificial intelligence model is learned using a learning algorithm using a plurality of learning data, thereby creating a predefined set of operation rules or an artificial intelligence model set to perform a desired characteristic (or purpose). This learning may be performed on the device itself on which the artificial intelligence according to the present disclosure is performed, or may be performed through a separate server and / or system. Examples of learning algorithms include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning.

[0034] An artificial intelligence model may be composed of multiple neural network layers. Each of the multiple neural network layers has multiple weight values, and performs neural network operations through operations between the operation results of the previous layer and the multiple weights. The multiple weights of the multiple neural network layers may be optimized based on the learning results of the artificial intelligence model. For example, the multiple weights may be updated so that the loss value or cost value obtained from the artificial intelligence model is reduced or minimized during the learning process. The artificial neural network may include a deep neural network (DNN), and examples thereof include, but are not limited to, a convolutional neural network (CNN), a deep neural network (DNN), a recurrent neural network (RNN), a restricted boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), or deep Q-networks.

[0035] According to an exemplary embodiment of the present disclosure, a processor can implement artificial intelligence. Artificial intelligence refers to a machine learning method based on an artificial neural network that imitates human neurons (biological neurons) to enable machines to learn. Artificial intelligence methodologies can be categorized into supervised learning, in which input data and output data are provided together as training data depending on the learning method, so that the solution (output data) to the problem (input data) is determined; unsupervised learning, in which only input data is provided without output data, so that the solution (output data) to the problem (input data) is not determined; and reinforcement learning, in which a reward (Reward) is provided from an external environment whenever an action (Action) is taken in the current state (State), and learning is performed in a direction to maximize this reward. In addition, artificial intelligence methodologies can be categorized according to the architecture of the learning model. The architectures of widely used deep learning technologies can be categorized into convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformers, and generative adversarial networks (GANs).

[0036] The present device and system may include an artificial intelligence model. The artificial intelligence model may be a single artificial intelligence model or may be implemented as multiple artificial intelligence models. The artificial intelligence model may be composed of a neural network (or artificial neural network) and may include statistical learning algorithms that mimic biological neurons in machine learning and cognitive science. A neural network may refer to a model in general that has problem-solving capabilities by changing the binding strength of synapses through learning, formed by artificial neurons (nodes) that form a network by combining synapses. The neurons of the neural network may include a combination of weights or biases. The neural network may include one or more layers composed of one or more neurons or nodes. For example, the device may include an input layer, a hidden layer, and an output layer. The neural network constituting the device can infer a desired result (output) from an arbitrary input (input) by changing the weights of neurons through learning.

[0037] The processor can create a neural network, train (or learn) a neural network, perform a calculation based on received input data, generate an information signal based on the calculation result, or retrain the neural network. The models of the neural network can include various types of models such as CNN (Convolution Neural Network) such as GoogleNet, AlexNet, VGG Network, R-CNN (Region with Convolution Neural Network), RPN (Region Proposal Network), RNN (Recurrent Neural Network), S-DNN (Stacking-based deep Neural Network), S-SDNN (State-Space Dynamic Neural Network), Deconvolution Network, DBN (Deep Belief Network), RBM (Restrcted Boltzman Machine), Fully Convolutional Network, LSTM (Long Short-Term Memory) Network, Classification Network, etc., but are not limited thereto. The processor can include one or more processors for performing calculations according to the models of the neural network. For example, the neural network can be a deep neural network. It may include a deep neural network.

[0038] Neural networks include CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), perceptron, multilayer perceptron, FF (Feed Forward), RBF (Radial Basis Network), DFF (Deep Feed Forward), LSTM (Long Short Term Memory), GRU (Gated Recurrent Unit), AE (Auto Encoder), VAE (Variational Auto) Encoder), DAE (Denoising Auto Encoder), SAE (Sparse Auto Encoder), MC (Markov Chain), HN (Hopfield Network), BM (Boltzmann Machine), RBM (Restricted Boltzmann Machine), DBN (Depp Belief Network), DCN (Deep Convolutional Network), DN (Deconvolutional Network), DCIGN (Deep Convolutional Inverse Graphics Network), Generative Adversarial Network (GAN), Liquid State Machine (LSM), Extreme Learning Machine (ELM), It will be understood by those skilled in the art that any neural network may be included, including but not limited to ESN (Echo State Network), DRN (Deep Residual Network), DNC (Differentiable Neural Computer), NTM (Neural Turning Machine), CN (Capsule Network), KN (Kohonen Network), and AN (Attention Network).

[0039] According to an exemplary embodiment of the present disclosure, the processor may be configured to perform a process for generating a CNN (Convolution Neural Network) such as GoogleNet, AlexNet, VGG Network, Region with Convolution Neural Network (R-CNN), Region Proposal Network (RPN), Recurrent Neural Network (RNN), Stacking-based deep Neural Network (S-DNN), State-Space Dynamic Neural Network (S-SDNN), Deconvolution Network, Deep Belief Network (DBN), Restrcted Boltzman Machine (RBM), Fully Convolutional Network, Long Short-Term Memory (LSTM) Network, Classification Network, Generative Modeling, eXplainable AI, Continual AI, Representation Learning, AI for Material Design, BERT, SP-BERT, MRC / QA for natural language processing, Text Analysis, Dialog System, GPT-3, GPT-4, Visual Analytics for vision processing, Visual Understanding, Video Synthesis, ResNet for data intelligence, Anomaly Detection, Prediction, Time-Series Forecasting, Various artificial intelligence structures and algorithms can be used, including but not limited to Optimization, Recommendation, and Data Creation.

[0040] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the attached drawings.

[0041] FIG. 1 is a block diagram illustrating an electronic device (10) to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure is applicable.

[0042] Referring to FIG. 1, an electronic device (10) to which a machine learning ensemble model for malaria diagnosis and detection can be applied may include an input / output unit (110), a communication unit (120), an identification unit (130), a database (140), and a processor (150). Hereinafter, the electronic device (10) is an electronic device that implements a machine learning ensemble model for malaria diagnosis and detection, and it is assumed that a malaria diagnosis and detection method to which a machine learning ensemble model is applied is implemented through the electronic device (10) to which the machine learning ensemble model for malaria diagnosis and detection can be applied.

[0043] The input / output unit (110) may be various interfaces or connection ports that receive user input or output information to the user. The input / output unit (110) may be divided into an input module and an output module.

[0044] The input module receives user input from the user. The input module is for inputting video information (or signals), audio information (or signals), data, or information input from the user, and may include at least one camera, at least one microphone, and at least one user input unit. Voice data or image data collected by the input unit may be analyzed and processed into user control commands.

[0045] User input can take many forms, including key input, touch input, and voice input. Examples of input modules that can receive such user input include, in addition to traditional keypads, keyboards, and mice, touch sensors that detect the user's touch, microphones that receive voice signals, cameras that recognize gestures through image recognition, proximity sensors such as light sensors or infrared sensors that detect user approach, motion sensors that recognize user movements such as acceleration sensors or gyro sensors, and a comprehensive concept that includes all types of input means that detect or receive various types of user input.

[0046] Here, the touch sensor may be implemented as a piezoelectric or electrostatic touch sensor that detects touch through a touch panel or touch film attached to the display panel, or an optical touch sensor that detects touch using an optical method. In addition, the input module may be implemented in the form of an input interface (such as a USB port or PS / 2 port) that connects an external input device that receives user input instead of a device that detects user input on its own.

[0047] An output module can output various types of information and provide them to the user. An output module is a comprehensive concept that includes a display that outputs images, a speaker (and / or an amplifier connected to it) that outputs sound, a haptic device that generates vibrations, and various other forms of output means. Additionally, an output module can be implemented in the form of a port-type output interface that connects the individual output means described above.

[0048] For example, an output module in the form of a display can display text, still images, and moving images. The display is a concept that broadly refers to an image display device that includes various forms of devices that can perform image output functions, such as a liquid crystal display (LCD), a light emitting diode (LED) display, an organic light emitting diode (OLED) display, a flat panel display (FPD), a transparent display, a curved display, a flexible display, a 3D display, a holographic display, a projector, and other devices. Such a display may be in the form of a touch display that is configured integrally with a touch sensor of the input module.

[0049] The communication unit (120) can communicate with an external device. Accordingly, the device can transmit and receive information with the external device through the communication unit. For example, the device can communicate with the external device using the communication unit so that information stored and generated within the electric vehicle charging management system can be shared. The communication unit (120) may include, for example, at least one of a wired communication module, a wireless communication module, a short-range communication module, and a location information module.

[0050] Here, communication, i.e., data transmission and reception, can be performed wired or wirelessly. To this end, the communication unit may be composed of a wired communication module that connects to the Internet, etc., via a LAN (Local Area Network), a mobile communication module that connects to a mobile communication network via a mobile communication base station to transmit and receive data, a short-range communication module that uses a WLAN (Wireless Local Area Network) series communication method such as Wi-Fi or a WPAN (Wireless Personal Area Network) series communication method such as Bluetooth or Zigbee, a satellite communication module that uses a GNSS (Global Navigation Satellite System) such as a GPS (Global Positioning System), or a combination thereof. The wireless communication technology used for communication may include NB-IoT (Narrowband Internet of Things) for low-power communication. At this time, for example, NB-IoT technology may be an example of LPWAN (Low Power Wide Area Network) technology and may be implemented with standards such as LTE Cat (category) NB1 and / or LTE Cat NB2, and is not limited to the above-described names. Additionally or alternatively, wireless communication technology implemented in wireless devices according to various embodiments may perform communication based on LTE-M technology. At this time, for example, LTE-M technology may be an example of LPWAN technology and may be called by various names such as eMTC (enhanced Machine Type Communication).For example, LTE-M technology can be implemented by at least one of various standards such as 1) LTE CAT 0, 2) LTE Cat M1, 3) LTE Cat M2, 4) LTE non-BL (non-Bandwidth Limited), 5) LTE-MTC, 6) LTE Machine Type Communication, and / or 7) LTE M, and is not limited to the above-described names. Additionally or alternatively, wireless communication technologies implemented in wireless devices according to various embodiments can include at least one of ZigBee, Bluetooth, and Low Power Wide Area Network (LPWAN) that take low-power communication into account, and are not limited to the above-described names. For example, ZigBee technology can create personal area networks (PANs) related to small / low-power digital communication based on various standards such as IEEE 802.15.4, and can be called by various names.

[0051] The wired communication module may include various wired communication modules such as a Local Area Network (LAN) module, a Wide Area Network (WAN) module, or a Value Added Network (VAN) module, as well as various cable communication modules such as a Universal Serial Bus (USB), a High Definition Multimedia Interface (HDMI), a Digital Visual Interface (DVI), RS-232 (recommended standard 232), power line communication, or plain old telephone service (POTS).

[0052] The wireless communication module may include a wireless communication module that supports various wireless communication methods such as GSM (global System for Mobile Communication), CDMA (Code Division Multiple Access), WCDMA (Wideband Code Division Multiple Access), UMTS (universal mobile telecommunications system), TDMA (Time Division Multiple Access), LTE (Long Term Evolution), 4G, 5G, and 6G, in addition to a WiFi module and a Wireless Broadband module.

[0053] The wireless communication module may include a wireless communication interface including an antenna and a transmitter for transmitting signals. Furthermore, the wireless communication module may further include a signal conversion module that modulates a digital control signal output from the control unit through the wireless communication interface into an analog wireless signal under the control of the control unit.

[0054] The wireless communication module may include a wireless communication interface including an antenna and a receiver for receiving signals. Furthermore, the wireless communication module may further include a signal conversion module for demodulating an analog wireless signal received through the wireless communication interface into a digital control signal.

[0055] The short-range communication module is for short-range communication, and is based on Bluetooth. TM), RFID (Radio Frequency Identification), infrared communication (Infrared Data Association; IrDA), UWB (Ultra Wideband), ZigBee, NFC (Near Field Communication), Wi-Fi (Wireless-Fidelity), Wi-Fi Direct, and Wireless USB (Wireless Universal Serial Bus) technologies can be used to support short-range communication.

[0056] The location information module is a module for obtaining the location (or current location) of the electronic device (10) according to the present disclosure, and representative examples thereof include a GPS (Global Positioning System) module or a WiFi (Wireless Fidelity) module. For example, by utilizing a GPS module, the location of the electronic device can be obtained using signals transmitted from GPS satellites. As another example, by utilizing a Wi-Fi module, the location of the electronic device (10) can be obtained based on information of a wireless access point (AP) that transmits or receives wireless signals with the Wi-Fi module. If necessary, the location information module may perform any function of other modules of the communication unit to obtain data regarding the location of the device as a substitute or in addition. The location information module is a module used to obtain the location (or current location) of the device, and is not limited to a module that directly calculates or obtains the location of the device.

[0057] The identification unit (130) may be a comprehensive concept that includes a camera that recognizes objects through image recognition, a detection sensor that detects the approach of an object, a touch sensor according to a user's input, and various other forms of identification / sensing means that detect or receive various forms of external input.

[0058] The camera processes image frames, such as still images or moving images, obtained by the image sensor in shooting mode. The processed image frames may be displayed on a display unit (or a screen of the electronic device of the present disclosure) or stored in memory.

[0059] Meanwhile, when there are multiple cameras, they can be arranged to form a matrix structure, and multiple image information having various angles or foci can be input through the cameras forming the matrix structure, and the cameras can also be arranged in a stereo structure to obtain left and right images for implementing a three-dimensional stereoscopic image.

[0060] The identification unit may be understood to be identical to the input module in the input / output unit (110) and / or may be understood to be separate from the input module. The identification unit (130) may further include one or more of a magnetic sensor, an acceleration sensor, a temperature / humidity sensor, an infrared sensor, a gyroscope sensor, a position sensor (e.g., GPS), a barometric pressure sensor, a proximity sensor, an RGB sensor (illuminance sensor), a radar sensor, a light sensor, and a current sensor, but is not limited thereto. Since the function of each sensor can be intuitively inferred from its name by those skilled in the art, a detailed description thereof will be omitted.

[0061] The database (140) can store various types of information. The database can store data temporarily or semi-permanently. For example, the database can store operating programs (OS: Operating System) for operating the device, data for hosting a website, programs for generating Braille, or data related to applications (e.g., web applications). Furthermore, the database can store modules in the form of computer code, as described above.

[0062] Examples of the database (140) may include a hard disk drive (HDD), a solid state drive (SSD), flash memory, read-only memory (ROM), random access memory (RAM), etc. Such a database may be provided as a built-in type or a detachable type.

[0063] The processor (150) controls the overall operation of the electronic device (10). To this end, the processor (150) can perform calculations and processing of various types of information and control the operation of components of the device. For example, the processor (150) can achieve rapid diagnosis and accurate detection when malaria infection is suspected through the first, second, and third models described below.

[0064] The processor (150) may be implemented as a computer or similar device, based on hardware, software, or a combination thereof. In terms of hardware, the processor (150) may be provided in the form of an electronic circuit that processes electrical signals to perform control functions, and in terms of software, the processor (150) may be provided in the form of a program that drives the hardware processor. Meanwhile, unless otherwise specified in the following description, the operation of the electronic device (10) may be interpreted as being performed under the control of the processor (150).

[0065] The processor (150) may be implemented as a memory storing data for an algorithm for controlling the operation of components within the device or a program reproducing the algorithm, and at least one subprocessor (not shown) that performs the aforementioned operations using the data stored in the memory. In this case, the memory and the processor may be implemented as separate chips. Alternatively, the memory and the processor may be implemented as a single chip.

[0066] In addition, the processor (150) can control any one or a combination of the components described above to implement various embodiments according to the present disclosure described in FIGS. 2 to 8 below on the device.

[0067] In Korea, vivax malaria, caused by the parasite Plasmodium vivax, is a frequent epidemic. Malaria infection can cause complications such as rupture of the spleen due to enlargement, and central nervous system abnormalities such as memory loss, convulsions, and schizophrenia. Therefore, prompt diagnosis and accurate detection of malaria infection are crucial factors that significantly impact treatment and patient prognosis.

[0068] While various diagnostic methods exist for diagnosing malaria infection in a patient, they can generally be determined based on whether malaria parasites are observed in the patient's peripheral blood. Therefore, the present disclosure proposes a method for detecting malaria parasites from a patient's peripheral blood image by incorporating object detection technology and a convolutional neural network (CNN), a deep learning technique specialized in image processing. Furthermore, the present disclosure proposes a method for detecting malaria infection with improved performance by utilizing not only image data but also blood test result data for malaria diagnosis in a supplementary / complementary manner using machine learning.

[0069] As a result, an image-based neural network model and an ensemble model that combines data machine learning can be built for malaria diagnosis.

[0070] According to an exemplary embodiment of the present disclosure, an electronic device (10) to which a machine learning ensemble model is applicable may be configured using a plurality of learning and / or prediction models including at least one neural network model.

[0071] For example, the electronic device (10) may include a first model learned based on learning images and a second model learned based on learning test data.

[0072] For example, the electronic device (10) may select or utilize a third model capable of combining the results of the image model and the data model. A model capable of selecting or combining multiple data results may also be referred to as an ensemble model. For example, the third model may be configured to output a probability of malaria infection based on a voting method for multiple statistical methodologies. However, the multiple models in this example are not intended to be limiting. The multiple learning and / or prediction models will be described in more detail with reference to FIGS. 2 and 7.

[0073] The processor (150) of the electronic device (10) as described above can use the first model to make a primary judgment on whether the blood image is infected with malaria, input the blood data to be tested into the second model, use the second model to make a secondary judgment on whether the blood data is infected with malaria, and use the third model to output the probability of malaria infection.

[0074] According to an electronic device and its operating method applicable to a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure, rapid diagnosis and accurate detection can be achieved when malaria infection is suspected.

[0075] In addition, according to an electronic device and an operating method thereof to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure is applicable, the relative number of malaria parasites according to the number of red blood cells can be counted through deep learning-based real-time object recognition technology, thereby enabling accurate reporting of the degree of malaria infection, and in order to supplement the detection of malaria parasites that deep learning cannot accurately detect, more accurate malaria parasite detection can be achieved by utilizing machine learning based on blood test results.

[0076] FIG. 2 is a block diagram illustrating in detail an artificial neural network applied to a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure.

[0077] Referring to FIG. 2, an electronic device (20) to which a machine learning ensemble model according to an exemplary embodiment of the present disclosure is applicable may include a first model (21), a second model (23), a third model (25), and an output model (27). However, this is merely an exemplary embodiment, and it does not exclude that at least one model may be omitted or other models may be included.

[0078] A malaria diagnosis and detection method based on a machine learning ensemble model, performed by an electronic device (20) according to an exemplary embodiment of the present disclosure, may include a step of inputting a blood image as a test subject into a first model trained with a neural network based on a training image, a step of first determining whether the blood image is infected with malaria using the trained first model, a step of inputting blood data as a test subject into a second model trained with a neural network based on the training test data, a step of second determining whether the blood data is infected with malaria using the trained second model, and a step of outputting a malaria infection probability using a third model that applies a voting method to a plurality of statistical methodologies.

[0079] According to an exemplary embodiment of the present disclosure, the first model (21) can be trained based on a training image. The training image may be a slide image composed by segmenting peripheral blood of a malaria patient.

[0080] According to an exemplary embodiment of the present disclosure, a total of 288 images for learning were used, including 168 malaria positive cases and 120 malaria negative cases, and 40 microscope slide images were used for each individual case, resulting in a total of 11,520 microscope slide images used as learning data.

[0081] According to an exemplary embodiment of the present disclosure, image data for individual malaria parasites may be used for training the first model (21). Each image data may be assigned a label for ground truth. For example, the first model (21) may be trained using label data for individual malaria parasites of each type, such as ring, trophozoite, gametocyte, schizont, and leukocyte and erythrocyte, as preprocessed training data. In a machine learning ensemble model-based malaria diagnosis and detection method performed by an electronic device (20) according to an exemplary embodiment of the present disclosure, the first model (21) may utilize the results of labeling of individual malaria parasites, including leukocytes, erythrocytes, ring, trophozoite, gametocyte, and schizont.

[0082] According to an exemplary embodiment of the present disclosure, for example, the first model (21) can augment data to improve model performance. For example, the first model (21) can enhance the amount of labeled data by utilizing the results of data augmentation performed using at least one of the horizontal flip method and the brightness range method, thereby improving learning and inference values.

[0083] According to an exemplary embodiment of the present disclosure, the first model (21) can be trained based on an artificial neural network after determining the learning target image size and batch size based on preprocessed data, labeled data, and augmented data. According to an exemplary embodiment of the present disclosure, Yolo-v5 among CNNs (Convolutional Neural Networks) was used for the first model. Specifically, the entire learning images were divided into a learning dataset and a test dataset at a ratio of 8:2 for all cases, and 20% of the learning dataset was used as validation data for Yolo-v5 to verify the model performance. The performance of the model will be described in more detail with reference to FIGS. 4 and 5.

[0084] According to an exemplary embodiment of the present disclosure, the first model (21) can output a prediction result of malaria infection by calculating a second coefficient for malaria parasites based on a first coefficient for red blood cells included in a blood image (IMG) that is a diagnostic target. Here, the first coefficient may have a statistical correlation with the second coefficient.

[0085] According to an exemplary embodiment of the present disclosure, the first model (21) can make a primary determination of malaria infection in a blood image by outputting a first coefficient value for red blood cells included in a blood image to be examined using a learned first neural network model, outputting a second coefficient value for malaria parasites based on the first coefficient value, and outputting a malaria infection prediction result based on the second coefficient value.

[0086] According to an exemplary embodiment of the present disclosure, the second model (23) can be trained based on learning test data. According to an exemplary embodiment of the present disclosure, for machine learning of the second model (23), various information including white blood cells, neutrophils (%), lymphocytes (%), monocytes (%), eosinophils (%), basophils (%), hemoglobin, red blood cells, platelet count (Plt), C-reactive protein (CRP), alanine aminotransferase, aspartate aminotransferase, total bilirubin, LDH, age, and gender can be collected from blood test data of 288 cases. Here, the number of missing values ​​in the total dataset is 422 (8.62%, 422 / 4896), and 4474 data were finally acquired.

[0087] According to an exemplary embodiment of the present disclosure, the second model (23) is configured to learn features with different importance levels, and can adopt at least one algorithm capable of modeling without replacing missing values. The second model (23) can perform a step of adopting at least one algorithm capable of modeling without replacing missing values ​​as part of a step of secondary determination of malaria infection in blood data using the learned second machine learning model.

[0088] For example, the second model (23) may use at least one of the eXtreme Gradient Boosting (XGBoost) model and the Light Gradient Boosting Model (LGBBM) model, but is not limited thereto. Models that are easy to model even if missing values ​​are left blank without separate preprocessing, such as eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LGBM), and CNN, may be used for learning the first and / or second models.

[0089] According to an exemplary embodiment of the present disclosure, for training the second model (23), a total of 288 cases, 168 malaria-positive and 120 malaria-negative, were divided into a training dataset and a test dataset at a ratio of 8:2, and the training dataset was trained and validated 10 times at a ratio of 9:1 (stratified K-fold validation). The performance of the second model (23) after training was verified using a separately classified 20% test dataset. Parameters affecting performance prediction were analyzed using feature importance techniques.

[0090] According to an exemplary embodiment of the present disclosure, the second model (23) may request supplementation if it determines that at least one detailed blood data, including CRP (C Reactive Protein) and Plt (Platelet), is missing from the blood data (BDAT). Here, CRP (C Reactive Protein) is a C-reactive protein, a type of protein synthesized in the liver and dissolved in the blood when inflammation occurs in the body, and Plt is the platelet count. Various abbreviations and values ​​of the experimental data will be described in more detail with reference to FIG. 4.

[0091] According to an exemplary embodiment of the present disclosure, the third model (25) can be configured to output a probability of infection (PROB) regarding malaria infection in a voting manner for multiple statistical methodologies.

[0092] According to an exemplary embodiment of the present disclosure, the third model (25) can function as a functional module that complements the malaria diagnosis results from the deep learning-based first model (21) and the machine learning-based second model (23). For example, the third model (35) can contribute to improving the area under the curve (AUROC), accuracy, precision, recall, and F1 score, which are evaluation criteria for malaria detection.

[0093] According to an exemplary embodiment of the present disclosure, the third model (25) can output a malaria infection verification result by re-judgement for a suspicious result value that falls below the diagnostic standard but is within the critical range among the predicted results of the first model. For example, the third model (25) can set a suspicious result value for an object that did not reach “0.8” in the first model (21) and was therefore not judged as “positive”, but has a value of “0.75”, which is within the range of -10% to +10% of the critical range, and re-judge this based on the blood test data of the second model (23), thereby verifying malaria infection. The third model (23) can improve the accuracy of malaria infection determination by verifying a suspicious value judged as “negative” in the first model (21) using the second model (23) and determining it as “positive.”

[0094] According to an exemplary embodiment of the present disclosure, the output model (27) may be configured to output a “positive” class, meaning malaria infection, or a “negative” class, meaning malaria non-infection, depending on whether the infection probability exceeds a threshold. Here, the “negative” class may be determined with greater dependency on the output of the first model (21) than on the output of the second model (23). Additionally, the “positive” class may be inferred from the voting results for an ensemble model that combines each of the plurality of statistical methodologies.

[0095] In a malaria diagnosis and detection method based on a machine learning ensemble model performed by an electronic device (20) according to an exemplary embodiment of the present disclosure, the step of outputting a malaria infection probability using a third model that applies a voting method to a plurality of statistical methodologies may include a step of outputting a “positive” class meaning malaria infection when the infection probability exceeds a threshold value, and a step of outputting a “negative” class meaning malaria non-infection when the infection probability is less than the threshold value.

[0096] According to an exemplary embodiment of the present disclosure, the output model (27) can determine whether malaria infection is "positive" if it determines that even one malaria parasite has been detected. Even if no malaria parasite has been detected, the output model (27) can determine whether malaria infection is present based on either or a combination of image data (IMG) and blood data (BDAT).

[0097] FIG. 3 is a diagram illustrating the performance of an electronic device (FIGS. 1, 10) to which a machine learning ensemble model for malaria diagnosis and detection can be applied according to an exemplary embodiment of the present disclosure. Referring to FIG. 3, the performance of the second model (23), a machine learning model based on blood test results and clinical information, is illustrated.

[0098] Figure 3 shows the malaria detection evaluation criteria, including AUROC, accuracy, precision, recall, F1 score, precision, overall average, and weighted average. The data in Figure 3 are briefly summarized in Table 1 below.

[0099] ModelAUROCAccuracyPrecisionRecallF1 scoreXGBoost0.99940.98280.98570.97920.9821LGBM1.00.98280.98000.98530.9823

[0100] Table 1 presents the performance of machine learning models based on blood test results and clinical information. In the second model (23), XGBoost achieved an accuracy of 0.9828, while LGBM achieved an accuracy of 0.9828. Meanwhile, the performance of the first model (21) is shown in Table 2.

[0101] ModelAccuracyPrecisionRecallF1 scoreYolo v50.8790.8291.00.907

[0102] Referring to Table 2, the first model (21) demonstrates the malaria classification performance of the YOLO v5 model based on peripheral blood slide images. Fifty-eight cases were used to validate the first model (21). The trained first model (21) accurately diagnosed 41 of the 58 patients as positive or negative, and diagnosed 7 of the 24 negative patients as positive.

[0103] In this case, the accuracy is 87.9% (=(34+17) / 58). The precision is 82.9% (=34 / (34+7)). The predictive value is the positive predicted value, meaning that the predicted positive result is actually positive.

[0104] Recall is sensitivity, where the denominator is (true positive + false negative; actual positive) and the numerator is (true positive), which in this case is 100% (=34 / 34).

[0105] The F1 score is the harmonic mean of precision and recall, and is scored as 0.907 (= 2 x (0.829 x 1) / (0.829+1)).

[0106] FIG. 4 is a diagram illustrating the performance of an electronic device (FIGS. 1 and 10) to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied. The performance of the second model (23), which has completed training, was verified using a separately classified 20% test dataset. As previously described, parameters that influenced performance prediction can be analyzed using feature importance techniques.

[0107] "Feature importance" is a technical methodology for interpreting variables in an ensemble model, and is a type of method for extracting the importance of variables when interpreting a model.

[0108] Figure 4 specifically shows the feature importance values ​​of XGboost. In the blood test data, white blood cells, neutrophils (%), lymphocytes (%), monocytes (%), eosinophils (%), basophils (%), hemoglobin, red blood cells, platelet count (Plt), C-reactive protein (CRP), alanine aminotransferase, aspartate aminotransferase, total bilirubin, LDH, age, and gender are shown, and among these, parameters with high values ​​have a significant impact on the implementation and interpretation of the second model (23).

[0109] According to Fig. 4, the feature importance of CRP (C Reactive Protein) is 0.289505, and the feature importance of Plt (Platelet) is 0.203164, confirming that they are relatively important compared to other variables. Therefore, the second model (23) may request supplementation if it is determined that at least one detailed blood data including CRP (C Reaction Protein) and Plt (Platelet) is missing.

[0110] FIG. 5 is a diagram showing the performance of an electronic device to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied.

[0111] Figure 5 illustrates the cell classification performance results of YOLO v5. Figure 5 shows the prediction rate and recall for each cell type / class, summarized in Table 3 below.

[0112] CellPrecisionRecallmAP (0.5)mAP (0.5-0.95)EpochAll0.8290.7000.7470.572100Gametocyte0.8050.9270.9160.671100Ring0.7490.7830.790 .571100Schizont1.00000.2420.218100Trophozoite0.6410.7880.8160.613100WBC0.9531.0000.9690.787100

[0113] FIGS. 6A to 6C are diagrams illustrating type-specific object recognition using a machine learning ensemble model for malaria diagnosis and detection according to exemplary embodiments of the present disclosure. FIG. 6A shows a real-world example of malaria detection using YOLO v5. In particular, FIG. 6A shows a detection result for a ring-type malaria parasite, indicating that the “ring” scoring score is “0.84” on the bounding box. FIG. 6B shows a real-world example of malaria detection using YOLO v5. In particular, FIG. 6B shows a detection result for a trophozoite-type malaria parasite, indicating that the “Trophozoite” scoring score is “0.87” on the bounding box.

[0114] Figure 6c shows a real-world example of malaria detection using YOLO v5. In particular, Figure 6c shows the detection results for gametocyte malaria parasites, indicating that the "Trophozoite" scoring score in the bounding box is "0.87."

[0115] FIG. 7 is a block diagram detailing a methodology applied to a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure.

[0116] According to an exemplary embodiment of the present disclosure, the third model (Fig. 2, 25) has a structure that determines a final prediction value by applying a voting method to the prediction values ​​of each of a decision tree, a K-nearest neighbor, and a logistic regression analysis. The prediction values ​​of each of the decision tree, the K-nearest neighbor, and the logistic regression analysis can be accumulated as a dataset.

[0117] According to an exemplary embodiment of the present disclosure, the third model (25) can apply a hard voting method. The results illustrating the application of the hard voting method to a total of three models (CNN, XGB, LGB) in the present disclosure are as shown in Table 4 below.

[0118] ModelAccuracyPrecisionRecallF1 scoreEnsemble1.01.01.01.0

[0119] In object recognition (CNN series Yolo v5), the accuracy was 0.881, and in machine learning, XGBoost and LGBM showed 0.9828 and 0.9828, respectively. Finally, the results of performing hard voting and classifying voting on two or more models out of three as the correct answer are as shown in Table 4, and the accuracy was 100%, showing accurate results in all cases. Ultimately, it was verified that the electronic device of the present disclosure to which the ensemble model is applied exhibited optimal performance. FIG. 8 is a block diagram illustrating an electronic device to which a machine learning ensemble model for malaria diagnosis and detection according to an exemplary embodiment of the present disclosure can be applied. The operating method of the electronic device to which the machine learning ensemble model according to the present disclosure can be applied, or the malaria infection diagnosis and detection method to which the machine learning ensemble model is applied, can be implemented in the form of a recording medium storing commands executable by a computer. The instructions may be stored in the form of program code, and when executed by a processor, may generate program modules to perform the operations of the disclosed embodiments. The recording medium may be implemented as a computer-readable recording medium.

[0120] Meanwhile, the disclosed embodiments may be implemented in the form of a recording medium storing computer-executable instructions. The instructions may be stored in the form of program code, and when executed by a processor, may generate program modules to perform the operations of the disclosed embodiments. The recording medium may be implemented as a computer-readable recording medium.

[0121] Computer-readable storage media include all types of storage media that store instructions that can be deciphered by a computer. Examples include read-only memory (ROM), random access memory (RAM), magnetic tape, magnetic disks, flash memory, and optical data storage devices.

[0122] An electronic device (FIGS. 1, 10) according to the present invention may correspond to a computing device (30), and the computing device (30) may include at least one processor (31), a computer-readable storage medium (32) including a program (33), and a communication bus (39). In addition, the computing device (30) may include one or more input / output interfaces (34) providing an interface for an input / output device (35), and one or more network communication interfaces (36).

[0123] Each component of the device according to the present invention may be included in one computing device (30) described above or may be implemented in a separate device.

[0124] The disclosed embodiments have been described with reference to the attached drawings as described above. Those skilled in the art will understand that the present disclosure can be implemented in forms other than the disclosed embodiments without altering the technical spirit or essential features of the present disclosure. The disclosed embodiments are illustrative and should not be construed as limiting.

Claims

1. In an electronic device to which a machine learning ensemble model can be applied, The first model is a neural network trained based on training images; A second model trained based on the training test data; A third model configured to output the probability of malaria infection by voting on multiple statistical methodologies; and An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, comprising a processor that inputs a blood image to be tested into the first model, performs a primary determination of malaria infection for the blood image using the first model, inputs blood data to be tested into the second model, performs a secondary determination of malaria infection for the blood data using the second model, and outputs the malaria infection probability using the third model.

2. In paragraph 1, The above first model is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, characterized in that a second count value for malaria parasites is calculated based on a first count value for red blood cells included in a blood image to be diagnosed, thereby outputting a prediction result for malaria infection.

3. In paragraph 1, The above learning image is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, characterized in that the electronic device is composed by segmenting (slide) peripheral blood of a malaria infected patient.

4. In paragraph 1, The second model above is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, characterized in that it adopts at least one algorithm that is configured to learn with different importance for each feature and can be modeled without replacing missing values.

5. In paragraph 4, The second model above is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, characterized by using at least one of an XGBoost (eXtreme Gradient Boosting) model and an LGBM (Light Gradient Boosting Model) model.

6. In paragraph 5, The second model above is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, characterized in that when at least one detailed blood data including CRP (C-Reactive Protein) and Plt (Platelet) is determined to be missing in the blood data to be examined, supplementation is requested.

7. In paragraph 1, The third model above is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, characterized in that it outputs a malaria infection verification result by re-judgement of a suspicious result value that falls below the diagnostic standard but is within the critical range among the predicted results of the first model.

8. In paragraph 1, The above first model is, An electronic device capable of applying a machine learning ensemble model for malaria diagnosis and detection, characterized in that it utilizes the results of labeling of individual malaria parasites (ring, trophozoite, gametocyte, schizont) and leukocytes and erythrocytes.

9. In paragraph 8, The above first model is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection can be applied, characterized in that the result of data augmentation is used using at least one of a horizontal flip methodology and a brightness range methodology.

10. In paragraph 1, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection is applicable, characterized in that it further includes an output module configured to output a positive class indicating malaria infection or a negative class indicating malaria non-infection depending on whether the infection probability exceeds a threshold value.

11. In paragraph 10, The above voice class is, An electronic device to which a machine learning ensemble model for malaria diagnosis and detection is applicable, characterized in that the output value of the first model has a greater dependency than the output value of the second model.

12. In paragraph 10, The above positive class is, An electronic device applicable to a machine learning ensemble model for malaria diagnosis and detection, characterized in that the result is inferred from a voting result for an ensemble model that combines each of the above multiple statistical methodologies.

13. A method for diagnosing and detecting malaria based on a machine learning ensemble model performed by an electronic device, A step of inputting a blood image to be examined into a first model trained by a neural network based on a learning image; A step of first determining whether the blood image is infected with malaria using the learned first model; A step of inputting blood data as a test subject into a second model learned based on learning test data; A step of making a secondary judgment on whether the blood data is infected with malaria using the learned second model; and A method for diagnosing and detecting malaria based on a machine learning ensemble model, comprising a step of outputting a malaria infection probability using a third model that applies a voting method to multiple statistical methodologies.

14. In paragraph 13, The first step of determining whether the blood image is infected with malaria using the first model learned above is as follows: A step of outputting a first count value for red blood cells included in a blood image that is a subject of examination; A step of outputting a second coefficient for malaria parasites based on the first coefficient; and A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that a malaria infection prediction result is output based on the second coefficient value.

15. In paragraph 13, The above learning image is, A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that the model is composed by segmenting (slide) the peripheral blood of a malaria infected patient.

16. In paragraph 13, The step of making a secondary judgment on whether or not the blood data is infected with malaria using the second model learned above is as follows: A method for diagnosing and detecting malaria based on a machine learning ensemble model, characterized in that it comprises a step of adopting at least one algorithm capable of modeling without replacing missing values.

17. In paragraph 16, The second model above is, A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that it is configured to learn with different importance for each feature and uses at least one of an XGBoost (eXtreme Gradient Boosting) model and an LGBM (Light Gradient Boosting Model) model.

18. In paragraph 17, The second model above is, A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that when at least one detailed blood data including CRP (C-Reactive Protein) and Plt (Platelet) is determined to be missing in the above blood data, supplementation is requested.

19. In paragraph 13, The step of outputting the probability of malaria infection using the third model that applies a voting method to multiple statistical methodologies is: A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that it includes a step of outputting a malaria infection verification result by re-judgement of a suspicious result value that falls below the diagnostic standard but is within the critical range among the predicted results of the first model.

20. In paragraph 13, The above first model is, A method for diagnosing and detecting malaria based on a machine learning ensemble model, characterized by utilizing the results of labeling of individual malaria parasites including white blood cells, red blood cells, and ring, trophozoite, gametocyte, and schizont.

21. In paragraph 20, The above first model is, A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that the result of data augmentation is used using at least one of a horizontal flip methodology and a brightness range methodology.

22. In paragraph 13, The step of outputting the probability of malaria infection using the third model that applies a voting method to the above multiple statistical methodologies is as follows: A step of outputting a “positive” class indicating malaria infection when the above infection probability exceeds a threshold value; and A method for diagnosing and detecting malaria based on a machine learning ensemble model, characterized in that it includes a step of outputting a “negative” class indicating non-malaria infection when the infection probability is less than a threshold value.

23. In paragraph 22, The above voice class is, A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that the output value of the first model has a greater dependency than the output value of the second model.

24. In paragraph 22, The above positive class is, A malaria diagnosis and detection method based on a machine learning ensemble model, characterized in that the result is inferred from a voting result for an ensemble model that combines each of the above multiple statistical methodologies.

Images