Cases in which personal terminals or servers are infected by ransomware are rapidly increasing. Ransomware uses a self-developed encryption module or combines existing symmetric key/public key encryption modules to illegally encrypt files stored in the victim system using a key known only to the attacker. Therefore, in order to decrypt it, it is necessary to know the value of the key used, and since the process of finding the decryption key takes a lot of time, financial costs are eventually paid. At this time, most of the ransomware malware is included in a hidden form in binary files, so when the program is executed, the user is infected with the malicious code without even knowing it. Therefore, in order to respond to ransomware attacks in the form of binary files, it is necessary to identify the encryption module used. Therefore, in this study, we developed a mechanism that can detect and identify by reverse analyzing the encryption module applied to the malicious code hidden in the binary file.
Deep learning with large amount of computations is difficult to implement on micro-sized IoT devices or moblie devices. Recently, lightweight deep learning technologies have been introduced to make sure that deep learning can be implemented even on small devices by reducing the amount of computation of the model. Quantization is one of lightweight techniques that can be efficiently used to reduce the memory and size of the model by expressing parameter values with continuous distribution as discrete values of fixed bits. However, the accuracy of the model is reduced due to discrete value representation in quantization. In this paper, we introduce various quantization techniques to correct the accuracy. We selected APoT and EWGS from existing quantization techniques, and comparatively analyzed the results through experimentations The selected techniques were trained and tested with CIFAR-10 or CIFAR-100 datasets in the ResNet model. We found out problems with them through experimental results analysis and presented directions for future research.
Recently, with the recent rapid development of memory technology, various types of memory are developed and are used to improve processing speed in data management systems. In particular, NAND flash memory is used as a main media for storing data in memory-based storage devices because it has a nonvolatile characteristic that it can maintain data even at the power off state. However, since the recently studied memory-based storage device consists of various types of memory such as MRAM and PRAM as well as NAND flash memory, research on memory management technology is needed to improve data processing performance and efficiency of media in a storage system composed of different types of memories. In this paper, we propose a memory mapping scheme thought technique for efficiently managing data in the storage device composed of various memories for data management. The proposed idea is a method of managing different memories using a single mapping table. This method can unify the address scheme of data and reduce the search cost of data stored in different memories for data tiering.
The brain-machine interface(BMI) is a next-generation interface that controls the device by decoding brain waves(also called Electroencephalogram, EEG), EEG is a electrical signal of nerve cell generated when the BMI user thinks of a command. The brain-machine interface can be applied to various smart devices, but complex computational process is required to decode the brain wave signal. Therefore, it is difficult to implement a brain-machine interface in an embedded system implemented in the form of an edge device. In this study, we proposed a new type of brain-machine interface system using IoT technology that only measures EEG at the edge device and stores and analyzes EEG data in the cloud computing. This system successfully performed quantitative EEG analysis for the brain-machine interface, and the whole data transmission time also showed a capable level of real-time processing.
고령화 사회로 인한 만성질환자의 증가로 그들의 질병 예방과 관리가 시급하다. 이러한 문제를 해결하기 위한생체인증 방법과 원격의료시스템들이 소개되고 있으나 의료정보와 개인인증의 보안성 문제를 해결하는데에는 어려움이있다. 스마트 헬스케어는 대상자의 개인 의료정보를 포함하고 있으므로 무엇보다 개인정보의 보안이 중요한 분야이다. 따라서 본 논문에서는 프라이빗 블록체인 환경에서 손목 밴드 형태의 스마트 웨어러블 디바이스 ECG와 얼굴 개인인증을 활용한 원격의료시스템을 제안하고자 한다. 이 시스템에서는 다양한 의료인과 전 지역 만성질환자를 대상으로 하였으며, 데이터의 무결성과 투명성을 높일 수 있는 프라이빗 블록체인을, 위변조가 어렵고 개인식별성이 높은 ECG와 얼굴인증을 활용하여 보안성과 신뢰성이 높일 수 있는 시스템을 구성하였다. 이를 통해 재가 만성질환자의 질병 예방과건강관리에 힘써 만성질환 관리의 효율성을 높이는데 기여하고자 한다.
Today, due to the increase in global population growth, the international community is discussing solving the food problem. The aquaculture industry is emerging as an alternative to solving the food problem. For the innovative growth of the aquaculture industry, smart fish farms that combine the fourth industrial technology are recently being distributed, and full-cycle digitalization is being promoted. Water quality sensors, which are important in the aquaculture industry, are electrochemical portable sensors that check water quality individually and intermittently, making it impossible to analyze and manage water quality in real time. Recently, optically-based monitoring sensors have been developed and applied, but the reliability of monitoring data cannot be guaranteed because the state information of the water quality sensor is unknown. Therefore, this paper proposes an algorithm representing self-diagnosis status such as Failure, Out of Specification, Maintenance Required, and Check Function based on monitoring data collected by water quality sensors to ensure data reliability.
The evolution of new digital technologies is progressing rapidly. In particular, many changes in software and artificial intelligence are progressing rapidly in the field of education. The Ministry of Education is planning an educational program by linking software and artificial intelligence regular curriculum. Before applying it to regular subjects, various software and artificial intelligence related experience camps are being promoted. This study aims to construct an educational model for software and artificial intelligence education programs for high school students based on new digital technology. By expanding and distributing software and artificial intelligence education, we aim to enhance the basic capabilities of software and artificial intelligence for high school students. I would like to define the concept of software and artificial intelligence in high school and propose a model that links software and artificial intelligence learning factors to the regular curriculum.
본 논문에서는 우산을 보다 쉽게 챙길 수 있도록 도와주는 사물인터넷 기반의 스마트 우산꽂이를 제안한다. 제안한 스마트 은 3가지 기능을 제공한다. 첫째, 기상 정보를 수신하여 정보에 따라 비가 온다면 우산의 손잡이를 노출시켜 사용자가 우산을 챙길 수 있도록 도와준다. 둘째, 스마트 우산꽂이는 열풍 시스템이 있어 빗물에 젖은 우산을 건조시킬 수 있다. 이를 통해 사용자는 빗물에 젖은 우산을 효율적으로 건조하고 쉽게 보관할 수 있도록 도와준다. 셋째, 앱(App)을 통해 스마트 우산꽂이의 현재 상태와 날씨, 빗물받이의 수위 등을 모니터링 할 수 있다. 제안한 스마트 우산꽂이는 비오는 날 우산을 챙겨갈 수 있도록 생활의 편의성을 제공할 것으로 기대한다.
In this paper, we propose an autonomous driving system using an end-to-end model to improve lane departure and misrecognition of traffic lights in a vision sensor-based system. End-to-end learning can be extended to a variety of environmental conditions. Driving data is collected using a model car based on a vision sensor. Using the collected data, it is composed of existing data and data with outlayers removed. A class was formed with camera image data as input data and speed and steering data as output data, and data learning was performed using an end-to-end model. The reliability of the trained model was verified. Apply the learned end-to-end model to the model car to predict the steering angle with image data. As a result of the learning of the model car, it can be seen that the model with the outlayer removed is improved than the existing model.
Recently, in order to store and manage big data, research and d evelopment of a high-performance storage system capable of stably accessing lar ge data have been actively conducted. In particular, storage systems in data centers and enterprise environments use large amounts of SSD (solid state disk) to manage large amounts of data. In general, SSD uses FTL(flash transfer layer) to hid e the characteristics of NAND flash memory, which is a medium, an d to efficiently manage data. However, FTL's algorithm has a limitation in using DRAM more to manage the location information of NAND where data is stored as the capacity of SSD increases. The refore, this paper introduces FTL policies that apply virtual memory to reduce DRAM resources use d in FTL. The virtual memory-based FTL policy proposed in this paper manages the map data by using LRU (least recently used) policy to load the mapping information of the recently used data into the DRAM space and store the previously used information in NAND. Finally, through experiments, perform ance and resource usage consumed during data write processing of virtual memory-based FTL and ge neral FTL are measured and analyzed.
Applying various association rule mining algorithms to the network intrusion detection task involves two critical issues: too large size of generated rule set which is hard to be utilized for IoT systems and hardness of control of false negative/positive rates. In this research, we propose an association rule mining algorithm based on the newly defined measures called coverage and exclusion. Coverage shows how frequently a pattern is discovered among the transactions of a class and exclusion does how frequently a pattern is not discovered in the transactions of the other classes. We compare our algorithm experimentally with the Apriori algorithm which is the most famous algorithm using the public dataset called KDDcup99. Compared to Apriori, the proposed algorithm reduces the resulting rule set size by up to 93.2 percent while keeping accuracy completely. The proposed algorithm also controls perfectly the false negative/positive rates of the generated rules by parameters. Therefore, network analysts can effectively apply the proposed association rule mining to the network intrusion detection task by solving two issues.
2007년 1월 국방부 유해발굴감식단이 창설되어 본격적인 유해발굴이 추진되었다. 현재 발굴 범위는 한국전쟁당시 치열한 전투가 벌어진 비무장지대 내의 화살머리고지까지 확장되어 진행되고 있다. 백마고지에서는 많은 유해 및유류품이 발굴되고 있다. 대부분 파손된 상태여서 정확한 형태를 복원하기는 어렵다. 이에 본 연구에서는 3D 스캔과3D 모델링을 통해 발굴된 유류품의 원형을 복원하고자 한다. 이러한 디지털 복원 방식은 그동안 수작업 방식의 단점을보완하는 원형 복원의 대안이 될 수 있다. 현재 문화유산 분야에서는 3D 기술을 사용한 각종 디지털 복원이 활발하다. 디지털 복원이 된 자료는 디지털 헤리티지의 기초자료로 활용될 수 있다. 이를 토대로 다양한 유해발굴 및 호국보훈관련 콘텐츠의 개발이 이루어질 수 있을 것이다. 더 나아가 발굴된 유해에 대한 디지털 휴먼 복원이 이루어진다면 전사자들의 생전 모습을 재현할 수 있을 것이다.
This paper is related to object segmentation using ESRGAN(Enhanced Super Resolution GAN) and SSS(Semantic Soft Segmentation). The segmentation performance of the object segmentation method using Mask R-CNN and SSS proposed by the research team in this paper is generally good, but the segmentation performance is poor when the size of the objects is relatively small. This paper is to solve these problems. The proposed method aims to improve segmentation performance of small objects by performing super-resolution through ESRGAN and then performing SSS when the size of an object detected through Mask R-CNN is below a certain threshold. According to the proposed method, it was confirmed that the segmentation characteristics of small-sized objects can be improved more effectively than the previous method.
This study seeks to find a way to induce users to expand their direct participation in sports through the acceptance of digital technology. From July 1 to August 30, 2022, a survey was conducted targeting home training users who applied the Internet of Things (IoT). 129 people participated in the survey through non-face-to-face self-administration method. For data processing, frequency analysis, exploratory factor analysis, reliability analysis, correlation analysis, multiple regression analysis, and 3-step mediation regression analysis were conducted using IBM's SPSS 21.0 program. The results of the study are as follows. First, in the relationship between the home training PPM model and direct participation in sports, ease appeared to have a mediating effect. In the factors of push, simple functionality showed a complete mediating effect, and inefficiency showed a partial mediating effect. Among pull factors, enjoyment and possibility of experience showed a complete mediating effect. In the mooring factors, individual innovativeness showed a complete mediating effect. Second, in the relationship between home training PPM model and direct participation in sports, usefulness showed a mediating effect. In the factors of push, simple functionality showed a complete mediating effect, and inefficiency showed a partial mediating effect. Among pull factors, enjoyment and possibility of experience showed a complete mediating effect. Among the mooring factors, individual innovativeness showed a partial mediating effect. Through this research, it is expected that the sports industry will contribute to the expansion of consumption expenditure and economic growth through the expansion of digital technologies such as NFT, Metaverse, and virtual/augmented reality.
While applying new digital technologies, interest in software and artificial intelligence is quite high. In particular, many changes are being made for the development of software and artificial intelligence in the field of education. From 2025, software and artificial intelligence-related curricula will be applied to public education in elementary, middle and high schools. The Ministry of Education is also conducting various camps to experience software and artificial intelligence in various ways in elementary, middle and high schools before they are applied to public education. Several platforms for experience camps related to software and artificial intelligence are also being used. In this study, we intend to increase the educational efficiency of the learning method for software and artificial intelligence to be developed in the future by designing a model for software and artificial intelligence experiential learning platforms.
Age of Information (AoI) has been introduced in wireless networked control systems (WNCSs) to guarantee timely status updates. In addition, as the edge computing (EC) architecture has been deployed in NCS, EC close to sensors can be exploited to collect status updates from sensors and provide control decisions to actuators. However, when lots of sensors simultaneously deliver status updates, EC can be overloaded, which cannot satisfy the AoI requirement. To mitigate this problem, this paper uses actuators with computing capability that can directly receive the status updates from sensors and determine the control decision without the help of EC. To analyze the AoI of the actuation update via EC or directly using actuators, this paper developed an analytic model based on timing diagrams. Extensive simulation results are included to verify the analytic model and to show the AoI with various settings.
This paper is related to performance improvement of Image-to-Image translation using Relativistic Average Patch GAN and Residual in Residual Dense Block. The purpose of this paper is to improve performance through technical improvements in three aspects to compensate for the shortcomings of the previous pix2pix, a type of Image-to-Image translation. First, unlike the previous pix2pix constructor, it enables deeper learning by using Residual in Residual Block in the part of encoding the input image. Second, since we use a loss function based on Relativistic Average Patch GAN to predict how real the original image is compared to the generated image, both of these images affect adversarial generative learning. Finally, the generator is pre-trained to prevent the discriminator from being learned prematurely. According to the proposed method, it was possible to generate images superior to the previous pix2pix by more than 13% on average at the aspect of FID.