Virtual autopsy detection device and method

Abstract

The present disclosure relates to a virtual autopsy detection device and method, which may: receive a two-dimensional image; preprocess the received two-dimensional image; learn the preprocessed two-dimensional image; generate a three-dimensional image and result data on the basis of the learned two-dimensional image; divide the generated three-dimensional image into a plurality of three-dimensional images; and display the plurality of divided three-dimensional images and the result data.

Description

Virtual autopsy detection device and method

[0001] The present disclosure relates to a virtual autopsy detection device and method. More specifically, it relates to a virtual autopsy detection device using an artificial intelligence learning model and a control method thereof.

[0002] In the legal system, a forensic autopsy is a medical procedure, similar to surgery, that dissects a corpse to determine the cause and nature of death. It analyzes the corpse for diseases and injuries to determine the cause of death. Based on scene records, investigation records, and statements from those involved in the case, it compares and evaluates forensic discrepancies between the process of death and the autopsy findings, providing legal evidence that aids in understanding the overall circumstances of the death.

[0003] In particular, corpses are prone to distortion and deformation of objective evidence due to continuous decomposition after death, and there is a vulnerability that interference may occur depending on the condition of the corpse before the autopsy, which may have a significant impact on the results of the subsequent autopsy. Therefore, a prompt autopsy must follow after securing sufficient on-site evidence at the scene of the crime.

[0004] At this point, the autopsy surgeon must not only establish a basic dissection plan for the corpse, but also thoroughly develop an autopsy strategy to match the vast amount of investigative and on-scene information. Therefore, an autopsy is a specialized medical procedure that requires both speed and objectivity, as well as precise dissection strategies, to prevent decomposition-related deformation. To this end, the surgeon may obtain additional assistance by obtaining a pre-autopsy CT scan of the corpse and sequentially reviewing the cross-sectional images before entering the autopsy room.

[0005] However, in the case of conventional technology, it was limited to simply taking a CT scan of the corpse before an autopsy, sequentially reviewing the cross-sectional images, and then proceeding with the autopsy, so there was a problem that users (autopsy doctors) felt inconvenienced because it was not directly helpful in autopsy practice that required speed and precision.

[0006] The purpose of the embodiment disclosed in the present disclosure is to provide a virtual autopsy detection device and method that learns abnormal structures in multiple layers of a corpse using artificial intelligence and provides result data and three-dimensional images for damaged parts of the corpse based on the learned abnormal structures.

[0007] The purpose of the embodiment disclosed in the present disclosure is to provide a virtual autopsy detection device and method capable of acquiring and managing full-body image data of a human body and developing a three-dimensional image processing technology.

[0008] The purpose of the embodiments disclosed in the present disclosure is to provide a virtual autopsy detection device and method that can develop an automatic zoning technology of human body structures using artificial intelligence and big data technology.

[0009] The purpose of the embodiments disclosed in the present disclosure is to provide a virtual autopsy detection device and method that can develop an extended reality-based virtual autopsy platform.

[0010] The problems to be solved by the present disclosure are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

[0011] A virtual autopsy detection device according to the present disclosure includes an input unit that receives a two-dimensional image; a control unit that preprocesses the received two-dimensional image, learns the preprocessed two-dimensional image, generates a three-dimensional image and result data based on the learned two-dimensional image, separates the generated three-dimensional image into a plurality of three-dimensional images, and displays the separated three-dimensional images and the result data; and a display that displays a graphic image according to a control command from the control unit.

[0012] In addition, a virtual autopsy detection method according to the present disclosure includes the steps of: receiving a two-dimensional image; preprocessing the received two-dimensional image; learning the preprocessed two-dimensional image; generating a three-dimensional image and result data based on the learned two-dimensional image; separating the generated three-dimensional image into a plurality of three-dimensional images; and displaying the separated plurality of three-dimensional images and the result data.

[0013] In addition, a computer program stored in a computer-readable recording medium may be further provided to execute a method for implementing the present disclosure.

[0014] In addition, a computer-readable recording medium recording a computer program for executing a method for implementing the present disclosure may be further provided.

[0015] According to the present invention, it is possible to learn abnormal structures in multiple layers of a corpse using artificial intelligence and provide result data and three-dimensional images for damaged parts of the corpse based on the learning, thereby enabling the autopsy results to be predicted in advance and the autopsy to be performed more accurately, thereby improving user convenience.

[0016] According to the present invention, by using a virtual autopsy detection method, forensic abnormal findings can be derived to provide forensic information for trauma analysis, thereby enabling autopsy results to be predicted in advance and autopsies to be performed more accurately, thereby improving user convenience.

[0017] According to the present invention, micro-damages that may be missed during an autopsy can be identified using AI, and the precision of the dissection and forensic focus of the reported area can be increased, and it has the advantage of being used as auxiliary data for diagnosis of the cause of death.

[0018] 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.

[0019] FIG. 1 is a diagram illustrating a configuration of a virtual autopsy detection device according to the present disclosure.

[0020] FIG. 2 is a diagram illustrating a flowchart of a virtual autopsy detection method according to the present disclosure.

[0021] FIG. 3 is a drawing illustrating an embodiment of an autopsy room to which a virtual autopsy detection method according to the present disclosure is applied.

[0022] FIG. 4 is a diagram illustrating a process of acquiring a two-dimensional CT image and a process of learning the normal structure of a body according to the present disclosure.

[0023] FIG. 5 is a diagram illustrating the creation of a three-dimensional image based on a two-dimensional CT image according to the present disclosure.

[0024] Figure 6 is a drawing showing the human body divided into skin, subcutaneous fat, muscle, and skeleton according to the present disclosure.

[0025] FIG. 7 is a diagram illustrating an example of a skin layer image and a subcutaneous fat image according to the present disclosure.

[0026] FIG. 8 is a diagram illustrating an embodiment of a muscle image and a skeletal image according to the present disclosure.

[0027] FIG. 9 is a diagram illustrating an embodiment of generating result data according to the present disclosure and displaying an abnormal structure when the result data is clicked.

[0028] FIG. 10 is a diagram illustrating annotation data corresponding to a human body according to the present disclosure.

[0029] FIG. 11 is a diagram illustrating displaying the results of a virtual autopsy according to the present disclosure in a two-dimensional mode and a three-dimensional mode.

[0030] FIG. 12 is a diagram illustrating a configuration of a virtual autopsy detection device according to the present disclosure.

[0031] 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 pertains 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.

[0032] Throughout the specification, when a part is said to be "connected" to another part, this includes not only direct connection but also indirect connection, and indirect connection includes connection via a wireless communication network.

[0033] 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.

[0034] 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.

[0035] The terms first, second, etc. are used to distinguish one component from another, and the components are not limited by the aforementioned terms.

[0036] Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0037] 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.

[0038] The operating principle and embodiments of the present invention will be described with reference to the attached drawings below.

[0039] The present invention can be implemented not only in a server system but also in various devices capable of performing computational processing and providing results to a user. For example, the present invention according to the present disclosure may include a computer, a server device, and a mobile terminal, or may be implemented in any one of the following forms.

[0040] Here, the computer may include, for example, a notebook, desktop, laptop, tablet PC, slate PC, etc. equipped with a web browser.

[0041] 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.

[0042] 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).

[0043] The related functions according to the present disclosure are operated through a processor and memory. The processor may be composed 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 DSP (Digital Signal Processor), a graphics-only processor such as a GPU or a VPU (Vision Processing Unit), or an artificial intelligence-only processor such as an NPU. One or more processors control the processing of input data according to predefined operation 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.

[0044] The predefined operation rules or artificial intelligence models are characterized by being created through learning. Here, being created through learning means that the basic artificial intelligence model is learned by a learning algorithm using a plurality of learning data, thereby creating a predefined operation rules or artificial intelligence model set to perform a desired characteristic (or purpose). This learning may be performed in 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 the learning algorithm include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning.

[0045] 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.

[0046] The processor can create a neural network, train (or learn) a neural network, perform computations based on received input data, and generate information signals based on the results of the computations, or retrain the neural network.

[0047] 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).

[0048] 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, including optimization, recommendation, and data creation, can be utilized, but are not limited thereto. Hereinafter, embodiments of the present disclosure will be described in detail with reference to the attached drawings.

[0049] FIG. 1 is a diagram illustrating a configuration of a virtual autopsy detection device according to the present disclosure.

[0050] Referring to FIG. 1, a virtual autopsy detection device (100) includes an input unit (110), a sensor unit (120), a control unit (130), a display (140), a memory (150), a communication unit (160), and a camera (170).

[0051] The input unit (110) receives a two-dimensional image.

[0052] The sensor unit (120) senses the state of a two-dimensional image.

[0053] The control unit (130) preprocesses the received two-dimensional image, learns the preprocessed two-dimensional image, generates a three-dimensional image and result data based on the learned two-dimensional image, separates the generated three-dimensional image into a plurality of three-dimensional images, and displays the separated three-dimensional images and the result data.

[0054] The control unit (130) learns the normal and abnormal structures of the body corresponding to the preprocessed two-dimensional image. A detailed description thereof is provided later in FIG. 4.

[0055] The control unit (130) divides the two-dimensional image into a first layer image, a second layer image, a third layer image, and a fourth layer image from the outside to the inside, and the first layer image corresponds to the skin of the body, the second layer image corresponds to the subcutaneous fat of the body, the third layer image corresponds to the muscles of the body, and the fourth layer image corresponds to the skeleton of the body, and learns the normal state of the skin, the normal state of the subcutaneous fat, the normal state of the muscles, and the normal state of the skeleton. A detailed description thereof will be described later with reference to FIG. 6.

[0056] The control unit (130) separates the three-dimensional image into a first layer image, a second layer image, a third layer image, and a fourth layer image.

[0057] Here, the first layer image corresponds to a skin layer image, the second layer image corresponds to a subcutaneous fat image, the third layer image corresponds to a muscle image, and the fourth layer image corresponds to a skeletal image. A detailed description thereof will be provided later in Figs. 7 and 8.

[0058] The control unit (130) determines that the skin corresponding to the first layer image is damaged when the first area in the first layer image exceeds the threshold range.

[0059] The control unit (130) determines that the subcutaneous fat corresponding to the second layer image is damaged when the second area in the second layer image exceeds the threshold range.

[0060] The control unit (130) determines that the muscle corresponding to the third layer image is damaged when the third area in the third layer image exceeds the threshold range.

[0061] The control unit (130) determines that the skeleton corresponding to the fourth layer image is fractured when the fourth region in the fourth layer image exceeds the threshold range. A detailed description of this will be described later in FIGS. 7 and 8.

[0062] The control unit (130) renders the learned two-dimensional CT image to generate a three-dimensional image. A detailed description thereof will be described later in FIG. 5.

[0063] The control unit (130) performs labeling on the normal structure and the abnormal structure to generate result data including annotation data. A detailed description thereof is provided later in FIG. 9.

[0064] When the control unit (130) receives an input from the user to select a first icon corresponding to the result data, it displays a screen corresponding to the first icon. A detailed description of this is provided later in FIG. 9.

[0065] The display (140) displays a graphic image according to a control command from the control unit (130).

[0066] The memory (150) stores inputs, results, and graphic images resulting from the execution of operations according to control commands from the control unit (130).

[0067] The control unit (150) stores learning data, 3D images, and result data in the memory (150).

[0068] The communication unit (160) transmits and receives data with an external device (200). Here, the external device (200) includes a smartphone, a PC, a laptop, a tablet PC, etc.

[0069] The camera (170) captures an image of the front according to a control command from the control unit (130).

[0070] However, the components illustrated in FIG. 1 are not essential for implementing a virtual autopsy detection device (100) according to the present disclosure, and thus the invention described in this specification may have more or fewer components than the components listed above.

[0071] The communication unit (160) may include one or more components that enable communication with an external device, and may include, for example, at least one of a broadcast reception module, a wired communication module, a wireless communication module, a short-range communication module, and a location information module.

[0072] The input unit (110) is for inputting video information (or signal), audio information (or signal), data, or information input from a 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 (110) may be analyzed and processed into a user control command.

[0073] The display (140) displays (outputs) information processed in the present invention. For example, the numerical anomaly detection device (100) may display execution screen information of a running application program (e.g., an application), or UI (User Interface) or GUI (Graphical User Interface) information based on such execution screen information.

[0074] The memory (150) can store data supporting various functions of the present invention, programs for the operation of the control unit, input / output data (e.g., music files, still images, moving images, etc.), and can store a plurality of application programs (or applications) driven by the present invention, data for the operation of the device, and commands. At least some of these application programs can be downloaded from an external server via wireless communication.

[0075] The memory (150) may include at least one type of storage medium among a flash memory type, a hard disk type, an SSD (Solid State Disk type), an SDD (Silicon Disk Drive type), a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, and an optical disk. In addition, the memory (150) is separate from the present invention, but may be a database connected by wire or wirelessly, and may be implemented as a database system.

[0076] The control unit (130) may be implemented with at least one core, a memory storing data for an algorithm for controlling the operation of components within the invention or a program reproducing the algorithm, and at least one processor (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.

[0077] In addition, the control unit (130) can control any one or a combination of the components discussed above in order to implement various embodiments according to the present invention described in FIGS. 2 to 12 below.

[0078] At least one component may be added or deleted to correspond to the performance of the components illustrated in Figure 1. Furthermore, it will be readily apparent to those skilled in the art that the relative positions of the components may be altered to correspond to the performance or structure of the system.

[0079] Meanwhile, each component illustrated in FIG. 1 refers to software and / or hardware components such as a Field Programmable Gate Array (FPGA) and an Application Specific Integrated Circuit (ASIC).

[0080] FIG. 2 is a diagram illustrating a flowchart of a virtual autopsy detection method according to the present disclosure.

[0081] The present invention is performed by a virtual autopsy detection device (100). In addition, the present invention can be performed by a control unit (130).

[0082] Referring to Fig. 2, a two-dimensional image is received (S210).

[0083] The received two-dimensional image is preprocessed (S220).

[0084] The preprocessed two-dimensional image is learned (S230).

[0085] A 3D image and result data are generated based on the learned 2D image (S240).

[0086] The generated 3D image is separated into multiple 3D images (S250).

[0087] The separated multiple three-dimensional images and the result data are displayed (S260).

[0088] FIG. 3 is a drawing illustrating an embodiment of an autopsy room to which a virtual autopsy detection method according to the present disclosure is applied.

[0089] The present invention is performed by a virtual autopsy detection device (100). In addition, the present invention can be performed by a control unit (130).

[0090] The control unit (130) displays autopsy result data and 3D images through the display (140).

[0091] Here, the autopsy results data includes AI Reporting results.

[0092] For example, AI Reporting results may include autopsy findings showing a punctured lung and a fractured upper right arm.

[0093] The 3D image can be an image converted from a 2D CT image of the skull of a corpse into a 3D image that visualizes it.

[0094] The control unit (130) can transmit and receive result data and 3D images to and from an external device (200). Here, the external device (200) includes a tablet PC.

[0095] Accordingly, the user (20) can predict the condition of the corpse and perform an autopsy by viewing multiple 3D images of the corpse and overlapping 3D images displayed on the tablet PC (200).

[0096] Explains applied anatomy.

[0097] Applied anatomy is a branch of medicine that applies anatomical knowledge to the diagnosis and treatment of disease. Only medical school anatomy departments can conduct applied anatomy research using donated cadavers. However, this research has currently been concentrated on dissection practice by medical students and surgical practice by residents, resulting in a constant oversupply of cadavers. Furthermore, physical limitations in space and equipment for dissections and autopsies inevitably limit the number of participants available for training.

[0098] This situation isn't unique to Korea; it's also occurring in developed countries, including the United States and Europe. Furthermore, anatomy education for healthcare professionals, including medical technicians, is inherently secondary to practical training, posing significant concerns that it could potentially lead to a decline in the quality of healthcare services in the long term.

[0099] To address this, we must regularly perform full-body CT scans on donated cadavers to continuously accumulate a vast amount of full-body imaging data, and, based on MOUs with world-class universities and multinational corporations, acquire automated technology for human body analysis by incorporating artificial intelligence and big data analysis, thereby promoting the development of digital analysis and morphological research methodology for full-body anatomy.

[0100] Describes the Virtual Autopsy Platform.

[0101] The augmented reality-based virtual autopsy platform developed in this research project is an integrated virtual autopsy solution that provides an autopsy simulation environment where the body of a deceased person can be virtually dissected before an actual autopsy, and can present artificial intelligence-based forensic abnormal findings and promote a comprehensive understanding of the process of death.

[0102] To this end, big data technology will be applied to present customized autopsy results based on full-body CT scans of deceased individuals. Furthermore, by providing simulation data on the mechanism of injury based on information from the scene and investigative records, this will significantly contribute to the development of autopsy strategies by forensic surgeons prior to actual autopsies and the accuracy of cause of death. Furthermore, the video data for the extended reality platform alone will not only have legal evidentiary value, but will also guide the focus on actual micro-autopsies based on the virtual autopsy results, securing broader evidentiary value. Furthermore, by presenting forensic abnormalities and their comprehensive interpretations through the platform, this platform will significantly enhance forensic understanding among relevant organizations, bereaved families, and other stakeholders, contributing significantly to the establishment of a social safety net.

[0103] In particular, by performing real-time automatic segmentation of CT scans of cadavers in various cases and creating a virtual dissection and virtual autopsy environment, it can greatly contribute to the development of medicine in various fields beyond just anatomy and forensic medicine, overcoming spatial limitations. In addition, if the target of the image is expanded to living patients, it will be possible to secure core technologies for the development of patient-tailored virtual medical services.

[0104] Explains national development strategies and policies.

[0105] First, let's explain cultural diversity.

[0106] Ethnic and cultural diversity is increasing worldwide, including in Korea, due to international migration and employment. In forensic medicine, cultural diversity not only affects ethnic differences in various diseases, but also manifests in differences in cognitive and behavioral patterns, leading to differences in the onset and mechanisms of injury.

[0107] This forensic awareness of cultural diversity needs to be actively applied to the interpretation of judicial cases, reflecting the current global trend of increasing multi-ethnic nations. Remaining confined to traditional, monocultural preconceptions can limit the scope of investigations when interpreting autopsies, hindering resolution. To address this, we need to collaborate with global research institutions to acquire medical imaging data from diverse ethnicities, analyze it, and simulate it within a standardized platform.

[0108] Second, we explain the globalization of forensic science.

[0109] Automated medical image analysis requires standardized automated segmentation technology for whole-body anatomy. Appropriate pixel distribution across the boundaries of these structures must be digitized using specific algorithms. This requires a platform that continuously improves these algorithms through big data technology, while AI can learn and optimize the classification of normal and abnormal structures.

[0110] The development of algorithms for 3D image processing and the development of forensic learning modules for artificial intelligence require proactively proposing standardized codes in line with the globalization trend. This will contribute to the globalization of forensic science and the incorporation of a virtual autopsy platform based on medical image processing into the Fourth Industrial Revolution.

[0111] To achieve this, a successful global virtual autopsy platform can be developed not only by leveraging the country's advanced technological capabilities but also by cooperating with the introduction of world-class artificial intelligence and big data operation technologies. Furthermore, exchanges with developing countries with inadequate forensic science capabilities can significantly contribute to the transfer of technological capabilities and the spread of global standardization.

[0112] Explains the lack of cadavers for dissection.

[0113] The shortage of cadavers for autopsy is a global phenomenon. In particular, acquiring the necessary forensic knowledge through hands-on dissections of various cases in legal practice and the current medical environment is extremely limited. This is due to the limited number of donated cadavers and the varying donation environments of each medical school. Forensic autopsies, particularly in the judicial system, are a crucial social decision-making process that determines the social responsibility of the deceased through irreproducible dissections. Therefore, the medical practice of autopsy requires the thorough preparation of an experienced autopsy surgeon who develops a systematic dissection strategy.

[0114] However, not all autopsy specialists have prior experience in anatomy or forensic medicine, and these specialists, including residents, medical students, and healthcare professionals in various fields, must prepare for a variety of situations with a limited number of donated cadavers or sampled human remains.

[0115] Explains the expansion of opportunities through virtual autopsy.

[0116] An extended reality (XR)-based virtual autopsy platform has no limitations regarding dissection or autopsy. The structural assessments and estimated causes of death suggested by AI learning can be used as valuable supplementary insights for expert decision-making. Beyond its educational potential, XR-based virtual autopsy platforms can also be used in legal practice to simulate the autopsy of a deceased individual in advance, develop hypotheses about the cause of death, and then conduct the autopsy to secure sufficient evidentiary power and solidify the logical connection between the results. This will significantly contribute to a comprehensive understanding of the deceased's death.

[0117] Explains how to use artificial intelligence and big data using full-body body scan images.

[0118] To date, technologies for handling whole-body CT images are very limited and have not yet been attempted. While partial simulations exist for specific medical procedures, these offer limited freedom based on a defined purpose and do not provide personalized solutions.

[0119] This is because opportunities to acquire large numbers of full-body images of people have been limited, and the need for technology capable of simultaneously processing and processing the massive amount of data in full-body images has not been raised. The virtual autopsy platform is an integrated 3D image processing solution that applies AI and big data technologies to process full-body images in real time, learn and supplement anatomical structure recognition, and furthermore, enables highly flexible dissections, making it applicable to forensic autopsy practice.

[0120] Explains the purpose of developing a virtual autopsy platform.

[0121] The purpose of this research and development project is to develop an augmented reality-based virtual autopsy platform that can be used in the specialized medical field of forensic anatomy. The platform learns anatomical knowledge about normal structures through manual and semi-automatic segmentation of the four major components of the human body, namely skin layer, subcutaneous fat layer, muscle layer, and bone layer, targeting whole-body CT. Based on this, the platform is equipped with an automatic labeling annotation tool for forensic abnormalities such as open wounds in the skin, subcutaneous hemorrhage, intermuscular hemorrhage, and fractures, which are representative damage findings in each layer. The platform is then capable of exploring forensic abnormalities and presenting 3D visualization reporting for the purpose of determining the cause of death tailored to the individual. This is an AI model-based platform.

[0122] By combining the 3D whole-body image processing technology acquired from the virtual autopsy platform with patient CT data, it will be possible to flexibly apply it to general medical fields such as the development of patient-tailored medical techniques and medical devices.

[0123] Describes an image management system for CT scans of the whole body.

[0124] First, over three years, we acquired full-body CT images from over 1,000 individuals. Each individual averaged 2,300 consecutively sectioned images at 0.7mm intervals, representing approximately 2.5GB of data. This data, combined with a Picture Archiving and Communication System (PACS), enabled continuous integration of 2,070,000 images (2,250GB) over three years. This image processing system facilitated image quality management and analysis. This also facilitated the establishment of an integrated management system for multi-ethnic whole-body CT data from the United States and other countries.

[0125] Second, the participation of hospital radiology technicians for medical image acquisition, medical school radiology experts for image quality improvement, image analysis experts for standardization and processing of digital medical images, and image network experts for management of large-capacity medical image storage and transmission is required.

[0126] Third, by utilizing imaging anatomy experts, we can manually segment approximately 2,070,000 CT cross-section images to separate the skin layer, subcutaneous fat layer, muscle layer, and bone layer of the whole body and label normal structures. We can also manually label forensic abnormalities such as open wounds in the skin, subcutaneous hemorrhage, intermuscular hemorrhage, and fractures, thereby building annotation data for abnormal structures.

[0127] Describes an image analysis system using artificial intelligence and big data.

[0128] First, by applying big data technology to the normal structures of the skin layer, subcutaneous fat layer, muscle layer, and bone layer of the whole body constructed through manual segmentation, and forensic abnormal findings of open wounds in the skin, subcutaneous hemorrhage, intermuscular hemorrhage, and fractures, a selective automatic distribution algorithm for the pixels constituting the boundaries of each structure is developed.

[0129] Second, we will develop an AI-based automatic labeling and annotation tool for forensic abnormalities in whole-body CT scans. This will allow us to develop a system that applies deep learning technology to acquire three-dimensional recognition capabilities and morphological heterogeneity for normal and abnormal anatomical structures by applying AI-supervised learning, a self-supervised learning method that learns by recognizing abnormal areas as normal.

[0130] Third, it is possible to develop a system with cognitive capabilities that can detect and explore forensic abnormalities and insight learning capabilities that can provide forensic information on trauma analysis, such as bleeding, hematoma, rupture, shear, fracture, dislocation, and tool damage.

[0131] A virtual autopsy platform using 3D visualization is described.

[0132] First, we will develop a virtual autopsy platform with a user interface similar to forensic autopsy practice and a highly flexible autopsy environment based on the autopsy surgeon's will. Based on full-body CT images of deceased individuals, the platform will enable highly flexible dissections and layer-by-layer structural exploration of specific areas to analyze the circumstances of the case. Furthermore, automatic 3D mesh generation and a tracking observation system will be provided to facilitate expanded educational applications as needed.

[0133] Second, as a platform tailored to the individual, it can provide a cyber communication space for communication and exchange of expert opinions among forensic experts around the world, transcending physical space.

[0134] Third, by building a virtual autopsy platform inspired by a telemedicine system, we can overcome religious and national restrictions on autopsies. By incorporating a digital autopsy system capable of transmitting and demonstrating data such as on-site 3D scans, we can explore ways to expand the utility of this integrated forensic autopsy solution.

[0135] The purpose of the present invention is as follows.

[0136] The purpose is to develop a platform for virtual autopsy based on extended reality.

[0137] The purpose is to develop a video analysis system utilizing AI and big data.

[0138] The goal is to provide a flexible, integrated solution suited to forensic medicine and various medical environments.

[0139] The effects of the present invention are as follows.

[0140] First, it is possible to secure the speed and accuracy of customized forensic autopsy of the deceased.

[0141] Second, the legal evidence can be supplemented through simulations of the estimated cause of death and the mechanism of damage.

[0142] Third, we can strengthen our focus on hypothesizing and micro-dissection of signs.

[0143] Fourth, it is possible to secure technology for rapid processing and management of large-capacity medical images.

[0144] Fifth, a substantial curriculum for anatomy and forensic medicine education can be established.

[0145] Sixth, it can help with dissection practice and increase education in the health and medical field.

[0146] Seventh, it is possible to develop patient-tailored medical devices.

[0147] Eighth, image analysis equipment can be developed based on AI and big data-based automatic zoning technology.

[0148] Other effects of the present invention are as follows.

[0149] First, a virtual autopsy platform can be established to improve education and work in forensic medicine-related institutions (medical schools, hospitals, National Forensic Service, National Police Agency, Prosecutors' Office, healthcare sector, etc.).

[0150] Second, it is possible to develop a human anatomy education system for everyone from the general public to experts and a social safety education module.

[0151] Third, image analysis equipment and medical devices can be developed using automatic segmentation technology of human anatomy structures.

[0152] The analysis of the interests of the present invention is as follows.

[0153] The development of an AR-based virtual autopsy platform can be utilized in practice and education at forensic institutions. In particular, it can be directly applied to the practice of performing autopsies on unnatural deaths by medical school professors, forensic professors, and forensic pathologists working at hospitals, hospitals, and the National Forensic Service, achieving diverse forensic outcomes. Furthermore, it can be utilized in forensic training and the development of field manual policies for the National Police Agency and Prosecutors' Office, where the importance of initial investigations is constantly emphasized, as well as for scientific investigators and emergency medical technicians dispatched to unnatural death scenes.

[0154] There are 40 medical schools nationwide, with approximately 100 anatomy professors, 40 forensic medicine professors, and 35 forensic medical examiners working at the National Forensic Service. The number of deaths witnessed or transferred to hospitals by emergency medical personnel accounts for approximately 70% of all unnatural deaths. Therefore, the need for forensic knowledge among hospital doctors is increasing day by day. Considering the number of medical professionals in hospitals nationwide, the implementation of this virtual autopsy platform in hospitals can be essential regardless of the size of the hospital.

[0155] There are approximately 2,000 field investigators and autopsy investigators affiliated with the National Police Agency nationwide, all of whom work on the front lines of forensic investigations. Considering the importance of initial investigations, training using a virtual autopsy platform and utilizing forensic autopsy simulations can be a significant turning point in solving death cases.

[0156] Other conflicts of interest analyses include:

[0157] To secure artificial intelligence and big data technologies for medical imaging, talent such as image analysis experts, image network experts, AI experts, and big data managers are needed. Conversely, after development, the training and education of these experts will play a crucial role.

[0158] Additionally, collaborative solutions from analytics consulting and big data management experts at U.S. universities and research institutes are needed.

[0159] Explains program management.

[0160] Explains the size of the virtual autopsy market.

[0161] The global forensic imaging market is estimated to be worth approximately $62.54 million. This growth is not only in absolute terms, but also reflects a trend of approximately 10.5% annual growth. This growth reflects the demand for rapid and accurate crime resolution. Forensic image analysis, particularly in the field of human imaging, often relies on simple analysis of raw data. Therefore, market demand is expected to grow further with the adoption of 3D stereoscopic imaging using virtual autopsy platforms.

[0162] Describes the size of the human anatomy education market.

[0163] In the era of the Fourth Industrial Revolution, the education industry market is expected to grow further, and the market for education utilizing virtual reality and augmented reality is particularly growing. The total global education market spending is expected to exceed $4 trillion, reaching approximately $8 trillion by 2025 and $10 trillion by 2030. In particular, creating human anatomy education programs using a customized augmented reality platform with an infinite number of examples is expected to differentiate itself from existing content in terms of both quality and quantity by developing a user interface that features two-way communication and a high degree of freedom in structural exploration.

[0164] Describes the derivative market (advanced medical imaging diagnostic equipment).

[0165] While medical imaging diagnostic services experienced a major leap forward with the advent of MRI and CT, they are now undergoing another paradigm shift, driven by AI technology. Specifically, the application of automatic recognition and recognition of human anatomy and automatic segmentation of internal structures can be applied to various medical imaging modalities, such as endoscopy and ultrasound. Furthermore, by training cross-sectional images containing lesions with AI, advanced medical imaging diagnostic devices can be developed that can identify lesions in a personalized manner and even suggest presumptive diagnoses, thereby improving the quality of healthcare services. Furthermore, the development of specialized image analysis systems for each body part is expected to enable the development of specialized medical devices.

[0166] Explains patent and medical device certification.

[0167] The present invention includes an extended reality-based virtual autopsy platform and operation technology, an artificial intelligence-based 3D automatic zoning technology for cross-section images, a technology for distributing boundary pixels of structures using big data, a technology for freely exploring interfaces for human body structures, development of haptic-based virtual human probes and incision hardware, and an educational virtual human anatomy image projection technology.

[0168] The effects of the present invention are explained.

[0169] The virtual autopsy platform, developed to simulate preliminary autopsies using CT cross-section images of actual corpses, will directly improve the work of forensic autopsies and forensic medicine, as well as enhance judicial services responsible for social safety. However, if the technologies contained in the platform are applied to education, healthcare, and social safety net construction, it will be sufficient to lead to the development of various forms of software, hardware, and medical devices.

[0170] In particular, the 3D automatic zoning and automatic labeling technology of the entire human body structure using big data and artificial intelligence can be expanded and reproduced to the field of future predictive image analysis by incorporating simulation of shape changes over time, and the technology for reproducing and predicting the formation mechanism of damage or external force is expected to play a responsible role in various insurance disputes and social recovery from disaster accidents.

[0171] FIG. 4 is a diagram illustrating a process of acquiring a two-dimensional CT image and a process of learning the normal structure of a body according to the present disclosure.

[0172] Figure 4 includes Figures 4(a) and 4(b).

[0173] Figure 4(a) is a diagram illustrating the process of acquiring a two-dimensional CT image from a corpse and the process of learning the normal structure of the body.

[0174] Figure 4(b) is a drawing showing a two-dimensional CT image obtained from a corpse.

[0175] Referring to Fig. 4(a), the control unit (130) acquires a two-dimensional CT image while moving the body (10) in the first direction while performing a photographing operation. Here, the two-dimensional image includes a two-dimensional CT image. Here, the body (10) can be a whole body or a part of the body.

[0176] As shown in Fig. 4(b), this is a two-dimensional CT image obtained from a corpse (10).

[0177] Here, the two-dimensional image includes a two-dimensional CT image and a two-dimensional MRI image.

[0178] The control unit (130) learns the normal and abnormal structures of the body corresponding to the preprocessed two-dimensional image.

[0179] The control unit (130) divides the two-dimensional image from the outside to the inside into a first layer image, a second layer image, a third layer image, and a fourth layer image.

[0180] The first layer image corresponds to the skin.

[0181] The second layer image corresponds to the subcutaneous fat.

[0182] The third layer images correspond to muscles.

[0183] The fourth layer image corresponds to the skeleton.

[0184] The control unit (130) learns the normal and abnormal structures of the skin.

[0185] The control unit (130) learns the normal and abnormal structures of subcutaneous fat.

[0186] The control unit (130) learns the normal and abnormal structures of the muscles.

[0187] The control unit (130) learns the normal and abnormal structures of the skeleton.

[0188] The control unit (130) learns normal and abnormal structures and reports on the learning results.

[0189] The control unit (130) determines that the lack of continuity of the skin is an abnormal structure.

[0190] In the case of subcutaneous fat, it detects areas that differ from the area of ​​the Housefield Unit of normal subcutaneous fat.

[0191] For muscles, it detects areas that differ from the area of ​​the House field unit of a normal muscle.

[0192] The Housefield Unit (HU) is the standard unit for CT scans. This unit can be used to identify objects in CT images. HU is based on water, with a reference value of 0. Its range is -1000 to 3000 and can be expressed as 12-bit data.

[0193] For example, for air, HU is -1000.

[0194] For water, HU is 0.

[0195] For soft tissue, HU is 100 to 300.

[0196] For the skeleton, the HU is 1800-1900 for the Cortcial (hard bones) and HU is 300-400 for the Cancellous (bone marrow tissue).

[0197] In the case of the skin, the control unit (130) determines that the part lacking continuity has an open damage.

[0198] In the case of bones, the control unit (130) determines that a part lacking continuity is a fracture. Or, it determines that a part where continuity is separated is a fracture.

[0199] In the case of subcutaneous fat, the control unit (130) determines that subcutaneous hemorrhage exists if an area where the HU value differs from the area specific to the subcutaneous fat is detected. In addition, if at least one of thickness and color changes is detected in the subcutaneous fat layer, the control unit determines that subcutaneous hemorrhage exists.

[0200] In the case of muscles, the control unit (130) determines that there is intermuscular bleeding when an area where the HU value is different from the area unique to the muscle space is detected.

[0201] FIG. 5 is a diagram illustrating the creation of a three-dimensional image based on a two-dimensional CT image according to the present disclosure.

[0202] Figure 5 includes Figures 5(a) and 5(b).

[0203] Figure 5(a) is a diagram illustrating a continuous arrangement of two-dimensional CT images acquired from a corpse. Here, the corpse image includes an image of the entire corpse (whole body).

[0204] Figure 5(b) is a diagram illustrating the creation of a 3D image based on a learned 2D CT image.

[0205] Specifically, the control unit (130) renders the learned two-dimensional CT images to generate three-dimensional images. Here, the number of two-dimensional CT images can be up to about 2,700.

[0206] Therefore, when creating a 3D image based on a 2D CT image, the 3D image can be created by appropriately adjusting the interval between the 2D CT images and then rendering.

[0207] Figure 6 is a drawing showing the human body divided into skin, subcutaneous fat, muscle, and skeleton according to the present disclosure.

[0208] As shown in Fig. 6, the control unit (130) divides the two-dimensional image (60) from the outside to the inside into a first layer image (61), a second layer image (62), a third layer image (63), and a fourth layer image (64).

[0209] The first layer image (61) corresponds to the skin of the body.

[0210] The second layer image (62) corresponds to the subcutaneous fat of the body.

[0211] The third layer image (63) corresponds to the muscles of the above body.

[0212] The fourth layer image (64) corresponds to the skeleton of the body.

[0213] The control unit (130) learns the normal state of the skin, the normal state of the subcutaneous fat, the normal state of the muscle, and the normal and abnormal states of the skeleton.

[0214] FIG. 7 is a diagram illustrating an example of a skin layer image and a subcutaneous fat image according to the present disclosure.

[0215] Figure 7 includes Figures 7(a) and 7(b).

[0216] Figure 7(a) is a drawing showing a first layer image corresponding to a skin layer image among three-dimensional images.

[0217] Figure 7(b) is a drawing showing a second layer image corresponding to a subcutaneous fat layer image among three-dimensional images.

[0218] The control unit (130) separates the three-dimensional image into a first layer image (710), a second layer image (720), a third layer image (810), and a fourth layer image (820).

[0219] Here, the first layer image (710) corresponds to a skin layer image, and the second layer image (720) corresponds to a subcutaneous fat image.

[0220] The control unit (130) determines that the skin corresponding to the first area is damaged when the first area in the first layer image (710) exceeds the threshold range.

[0221] For example, if a specific area of ​​the first layer image lacks continuity or is interrupted, the skin corresponding to the specific area is determined to be open. Here, the threshold value refers to a value within the range where continuity of the specific area is guaranteed.

[0222] The control unit (130) determines that the subcutaneous fat corresponding to the second area is damaged when the second area in the second layer image exceeds the threshold range.

[0223] For example, if a specific area of ​​the second layer image exceeds the HU value specific to subcutaneous fat, it is determined that there is subcutaneous hemorrhage in the subcutaneous fat corresponding to the specific area. Here, the threshold value refers to the HU value specific to subcutaneous fat.

[0224] The control unit (130) displays a plurality of separated three-dimensional images in different colors.

[0225] For example, the first layer image (710) is displayed in blue, the second layer image (720) in red, the third layer image (810) in orange, and the fourth layer image (820) in white.

[0226] FIG. 8 is a diagram illustrating an embodiment of a muscle image and a skeletal image according to the present disclosure.

[0227] Figure 8 includes Figures 8(a) and 8(b).

[0228] Figure 8(a) is a drawing showing a third layer image corresponding to a muscle layer image among three-dimensional images.

[0229] Figure 8(b) is a drawing showing a fourth layer image corresponding to a skeleton image among three-dimensional images.

[0230] The control unit (130) separates the three-dimensional image into a first layer image (710), a second layer image (720), a third layer image (810), and a fourth layer image (820).

[0231] The third layer image (810) corresponds to a muscle image.

[0232] The fourth layer image (820) corresponds to the skeleton image.

[0233] The control unit (130) determines that the muscle corresponding to the third layer image (810) is damaged when the third area in the third layer image (810) exceeds the threshold range.

[0234] For example, if a specific area of ​​the third layer image (810) exceeds the muscle-specific HU value, it is determined that the muscle corresponding to the specific area is damaged. Here, the threshold value means the muscle-specific HU value.

[0235] The control unit (130) determines that the skeleton corresponding to the fourth layer image (820) is fractured when the fourth area in the fourth layer image (820) exceeds the threshold range.

[0236] For example, if a specific area of ​​the fourth layer image (820) lacks continuity or if the continuity of a specific area is broken, it is determined that the skeleton corresponding to the specific area is fractured. Here, the threshold value means a value within a range where continuity of the specific area is guaranteed.

[0237] FIG. 9 is a diagram illustrating an embodiment of generating result data according to the present disclosure and displaying an abnormal structure when the result data is clicked.

[0238] Figure 9 includes Figures 9(a) and 9(b).

[0239] Figure 9(a) is a diagram showing the result data and body image.

[0240] Figure 9(b) is a diagram illustrating an abnormal structure corresponding to the result data.

[0241] As shown in Fig. 9(a), the control unit (130) performs labeling on the normal structure and the abnormal structure to generate result data including annotation data.

[0242] The control unit (130) generates result data including an automatic inspection report on abnormal parts based on a 2D CT image and a 3D visualization image, and displays the same.

[0243] Here, the automated inspection report is a report on the suspected part.

[0244] For example, if there is a part of the muscle in the upper part of the left arm that is identified as muscle hemorrhage and result data including the corresponding report content is clicked, the control unit (130) displays an abnormal structure corresponding to the muscle hemorrhage as a three-dimensional image.

[0245] For example, the control unit (130) performs labeling of normal and abnormal structures of a corpse (10), generates result data (911) including annotation data, and displays the generated result data (911).

[0246] For example, the result data (911) includes Reporting a left humerus fracture.

[0247] The control unit (130) overlays a graphic image corresponding to the result data on the corpse image and displays it.

[0248] For example, a graphic image (912) corresponding to a fractured left arm bone can be overlaid and displayed on top of a full body image (910).

[0249] As shown in Fig. 9(b), when the control unit (130) receives an input from the user to select a first icon corresponding to the result data, it displays a screen corresponding to the first icon.

[0250] For example, when an input is received from a user for selecting an icon (911) corresponding to a left arm bone fracture, a screen (920) corresponding to a left arm bone fracture is displayed.

[0251] Additionally, the control unit (130) overlays and displays an indicator (922) indicating a fractured left arm bone on a two-dimensional screen (920).

[0252] When the control unit (130) receives a screen enlargement input from the user, it enlarges and displays the two-dimensional screen (920). At this time, the resolution of the two-dimensional screen (920) is increased.

[0253] Resolution can be 360p, 480p, 720p, 1080p, 1440p, 2160p, and the resolution increases as the screen gets larger.

[0254] According to the present invention, by performing labeling for normal structures and abnormal structures, result data including annotation data is generated, the generated result data is displayed, and when the result data is clicked, the abnormal structure is displayed, so that the user can intuitively know the result data and the problematic part, thereby improving user convenience.

[0255] FIG. 10 is a diagram illustrating annotation data corresponding to a human body according to the present disclosure.

[0256] Referring to Figure 10, annotation data is written in the order of Code, Region, Part, Name, and Abbr.

[0257] Let me explain with an example.

[0258] For Code 1-H-1, Region is Head, Part is Skin, and Name is H-Skin. Abbreviation is HS.

[0259] For Code 1-H-2, Region is Head, Part is Subcutaneous Q, and Name is H-SubQ. Abbreviation is HQ.

[0260] For Code 1-H-3, the Region is Head, the Part is Muscle, and the Name is H-Muscle. The abbreviation is HM.

[0261] For Code 1-H-4, Region is Head, Part is Bone, and Name is Frontal. Abbreviation is HF.

[0262] FIG. 11 is a diagram illustrating displaying the results of a virtual autopsy according to the present disclosure in a two-dimensional mode and a three-dimensional mode.

[0263] Figure 11 includes Figures 11(a) and 11(b).

[0264] Figure 11(a) is a diagram illustrating a two-dimensional mode and a three-dimensional mode.

[0265] Figure 11(b) is a diagram showing the results of a two-dimensional mode autopsy and a three-dimensional mode autopsy.

[0266] As shown in Fig. 11(a), the control unit (130) displays a two-dimensional mode (1112) and a three-dimensional mode (1114) when displaying the virtual autopsy results.

[0267] When the control unit (130) receives an input from the user (10) to select a two-dimensional mode (1112), it displays a virtual autopsy result (1120) corresponding to the two-dimensional mode, as shown in Fig. 11(b).

[0268] When the control unit (130) receives an input from the user (10) to select a 3D mode (1114), it displays a virtual autopsy result (1130) corresponding to the 3D mode, as shown in Fig. 11(b).

[0269] Next, we will explain the indicators that indicate the location of the skull fracture.

[0270] The control unit (130) overlays an indicator (1122) indicating a skull fracture area on the virtual autopsy results (1120) corresponding to the two-dimensional mode and displays them.

[0271] The control unit (130) displays an indicator (1132) indicating a skull fracture area by overlaying it on the virtual autopsy results (1130) corresponding to the 3D mode.

[0272] According to the present invention, two-dimensional mode and three-dimensional mode can be displayed, and by clicking on each mode, the virtual autopsy results can be displayed, so that the user can intuitively know the virtual autopsy results and the problematic parts, thereby improving user convenience.

[0273] FIG. 12 is a diagram illustrating a configuration of a virtual autopsy detection device according to the present disclosure.

[0274] Referring to FIG. 12, the virtual autopsy detection device (100) includes a device (1600). The device (1600) may include a memory (1602), a processor (1603), a transceiver (1604), and a peripheral device (1601). In addition, as an example, the device (1600) may further include other configurations and is not limited to the above-described embodiment.

[0275] More specifically, the device (1600) of FIG. 12 may be an exemplary hardware / software architecture, such as an NDN device, an NDN server, or a content router. As an example, the memory (1602) may be non-removable memory or removable memory. Furthermore, as an example, the peripheral device (1601) may include a display, GPS, or other peripheral devices, and is not limited to the above-described embodiment.

[0276] In addition, as an example, the above-described device (1600) may include a communication circuit such as the transceiver (1604), and may perform communication with an external device based thereon.

[0277] Additionally, as an example, the processor (1603) may be at least one of a general-purpose processor, a digital signal processor (DSP), a DSP core, a controller, a microcontroller, ASICs (Application Specific Integrated Circuits), FPGA (Field Programmable Gate Array) circuits, any other type of integrated circuit (IC), and one or more microprocessors associated with a state machine. In other words, it may be a hardware / software configuration that performs a control role for controlling the above-described device (1600).

[0278] At this time, the processor (1603) may execute computer-executable instructions stored in the memory (1602) to perform various essential functions of virtual autopsy detection. For example, the processor (1603) may control at least one of signal coding, data processing, power control, input / output processing, and communication operations. In addition, the processor (1603) may control the physical layer, the MAC layer, and the application layers. In addition, for example, the processor (1603) may perform authentication and security procedures at the access layer and / or the application layer, and is not limited to the above-described embodiment.

[0279] For example, the processor (1603) can communicate with other devices through the transceiver (1604).

[0280] For example, the processor (1603) can control the execution of computer-executable instructions to detect document similarity. That is, the communication performed in the present invention can be controlled. For example, other nodes may be NDN servers, content routers, and other devices. For example, the transceiver (1604) can transmit RF signals via an antenna and transmit signals based on various communication networks.

[0281] In addition, as an example, MIMO technology, beamforming, etc. can be applied as antenna technology, and are not limited to the above-described embodiment. In addition, the signal transmitted and received through the transceiver (1604) can be modulated and demodulated and controlled by the processor (1603), and are not limited to the above-described embodiment.

[0282] The various embodiments of the present disclosure are not intended to list all possible combinations but rather to illustrate representative aspects of the present disclosure, and the matters described in the various embodiments may be applied independently or in combinations of two or more.

Claims

1. Display; an input module for receiving a two-dimensional image; and Preprocessing the received two-dimensional image, Learning the above preprocessed 2D image, Generate a 3D image and result data based on the learned 2D image, A virtual autopsy detection device comprising a processor that separates the generated three-dimensional image into a plurality of three-dimensional images and displays the separated three-dimensional images and the result data.

2. In the first paragraph, the processor, A virtual autopsy detection device that learns normal and abnormal structures of a body corresponding to the preprocessed two-dimensional image.

3. In the second paragraph, the processor, A virtual autopsy detection device that generates result data including annotation data by performing labeling on the above normal structure and the above abnormal structure.

4. In paragraph 2, The above processor, Divide the above two-dimensional image into a first layer image, a second layer image, a third layer image, and a fourth layer image from the outside to the inside, Correspond the above first layer image to the skin of the body, The second layer image is corresponded to the subcutaneous fat of the body, Correspond the above third layer image to the muscles of the body, Correspond the above fourth layer image to the skeleton of the body, A virtual autopsy detection device that learns the normal state of the skin, the normal state of the subcutaneous fat, the normal state of the muscle, and the normal state of the skeleton.

5. In the first paragraph, the processor, Separate the above 3D image into a first layer image, a second layer image, a third layer image, and a fourth layer image, Correspond the above first layer image to the skin layer image, Correspond the above second layer image to the subcutaneous fat image, Correspond the above third layer image to the muscle image, Corresponding the above 4th layer image to the skeleton image, Virtual autopsy detection device.

6. In the fifth paragraph, the processor, If the first region in the first layer image exceeds the threshold range, it is determined that the skin corresponding to the first region is damaged, If the second region in the second layer image exceeds the threshold range, it is determined that the subcutaneous fat corresponding to the second region is damaged, If the third region in the third layer image exceeds the threshold range, it is determined that the muscle corresponding to the third region is damaged, If the fourth region in the fourth layer image exceeds the threshold range, it is determined that the bone corresponding to the fourth region is fractured. Virtual autopsy detection device.

7. In the first paragraph, the processor, Rendering the learned 2D CT image to create a 3D image, When receiving an input from a user to select a first icon corresponding to the result data, display a screen corresponding to the first icon, Displaying the separated multiple three-dimensional images in different colors, A virtual autopsy detection device that preprocesses a plurality of consecutive two-dimensional images by screening them at predetermined intervals.

8. In a virtual autopsy detection method performed by a device, A step of receiving a two-dimensional image; A step of preprocessing the received two-dimensional image; A step of learning the preprocessed two-dimensional image; A step of generating a three-dimensional image and result data based on the learned two-dimensional image; A step of separating the generated 3D image into multiple 3D images; and A virtual autopsy detection method comprising a step of displaying the separated plurality of three-dimensional images and the result data.

9. In paragraph 8, A virtual autopsy detection method further comprising a step of learning normal and abnormal structures of a body corresponding to the preprocessed two-dimensional image.

10. In paragraph 9, A virtual autopsy detection method further comprising a step of generating result data including annotation data by performing labeling on the normal structure and the abnormal structure.

11. In paragraph 9, A step of dividing the above two-dimensional image into a first layer image, a second layer image, a third layer image, and a fourth layer image from the outside to the inside; A step of corresponding the first layer image to the skin of the body; A step of corresponding the second layer image to the subcutaneous fat of the body; A step of corresponding the third layer image to the muscles of the body; A step of corresponding the fourth layer image to the skeleton of the body; and A virtual autopsy detection method further comprising a step of learning the normal state of the skin, the normal state of the subcutaneous fat, the normal state of the muscle, and the normal state of the skeleton.

12. In paragraph 8, A step of dividing the above three-dimensional image into a first layer image, a second layer image, a third layer image, and a fourth layer image; A step of corresponding the first layer image to a skin layer image; A step of corresponding the second layer image to the subcutaneous fat image; A step of corresponding the third layer image to a muscle image; and A virtual autopsy detection method further comprising the step of corresponding the fourth layer image to a skeletal image.

13. In paragraph 12, A step of determining that the skin corresponding to the first region is damaged when the first region in the first layer image exceeds the threshold range; A step of determining that the subcutaneous fat corresponding to the second region is damaged when the second region in the second layer image exceeds the threshold range; A step of determining that a muscle corresponding to the third region is damaged when the third region in the third layer image exceeds the threshold range; and Further comprising a step of determining that a bone corresponding to the fourth region is fractured when the fourth region in the fourth layer image exceeds a threshold range. A method for detecting virtual autopsy.

14. In paragraph 8, A step of rendering the learned two-dimensional CT image to generate a three-dimensional image; When receiving an input from a user for selecting a first icon corresponding to the result data, a step of displaying a screen corresponding to the first icon; A step of displaying the separated plurality of three-dimensional images in different colors; and A virtual autopsy detection method further comprising a step of preprocessing by screening the plurality of consecutive two-dimensional images at predetermined intervals.

15. A computer-readable storage medium having recorded thereon a computer program for performing the method of Article 8.

Images