• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.5
• /
• pp.193-204
• /
• 2019

The explosion of data due to the improvement of sensor technology and computing performance has become the basis for analyzing the situation in the industrial fields, and various attempts to detect events based on such data are increasing recently. In particular, sound signals collected from sensors are used as important information to classify events in various application fields as an advantage of efficiently collecting field information at a relatively low cost. However, the performance of sound-event classification in the field cannot be guaranteed if noise can not be removed. That is, in order to implement a system that can be practically applied, robust performance should be guaranteed even in various noise conditions. In this study, we propose a system that can classify the sound event after generating the enhanced sound signal based on the deep learning algorithm. Especially, to remove noise from the sound signal itself, the enhanced sound data against the noise is generated using SEGAN applied to the GAN with a VAE technique. Then, an end-to-end based sound-event classification system is designed to classify the sound events using the enhanced sound signal as input data of CNN structure without a data conversion process. The performance of the proposed method was verified experimentally using sound data obtained from the industrial field, and the f1 score of 99.29% (railway industry) and 97.80% (livestock industry) was confirmed.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.9
• /
• pp.404-412
• /
• 2024

The adoption of cloud environments has revolutionized application deployment and management, with microservices architecture and container technology serving as key enablers of this transformation. However, these advancements have introduced new challenges, particularly the necessity to precisely understand service interactions and conduct detailed analyses of internal processes within complex service environments such as microservices. Traditional monitoring techniques have proven inadequate in effectively analyzing these complex environments, leading to increased interest in eBPF (extended Berkeley Packet Filter) technology as a solution. eBPF is a powerful tool capable of real-time event collection and analysis within the Linux kernel, enabling the monitoring of various events, including file system activities within the kernel space. This paper proposes a container activity analysis system based on eBPF, which monitors events occurring in the kernel space of both containers and host systems in real-time and analyzes the collected data. Furthermore, this paper conducts a comparative analysis of prominent eBPF-based container monitoring systems (Tetragon, Falco, and Tracee), focusing on aspects such as event detection methods, default policy application, event type identification, and system call blocking and alert generation. Through this evaluation, the paper identifies the strengths and weaknesses of each system and determines the necessary features for effective container process monitoring and restriction. In addition, the proposed system is evaluated in terms of container metadata collection, internal activity monitoring, and system metadata integration, and the effectiveness and future potential of eBPF-based monitoring systems.

• Korean Journal of Animal Reproduction
• /
• v.23 no.4
• /
• pp.303-311
• /
• 1999

Nuclear DNA was isolated from the sperm cells representing genetic characteristics and genomic polymorphisms of rainbow trout by polymerase chain reaction(PCR) amplification of DNA using arbitrary primers. Genomic DNA fingerprints were generated from rainbow trout sperm DNA by polymerase chain reaction amplification using 20 arbitrary decamers as primers. Out of these primers, 4 generated 17 highly reproducible RAPD markers, producing almost six polymorphic bands per primers. Four of 6 primers tested generated amplified fragments which were polymorphic between different individuals. Polymorphic DNA fragments were reproducibly amplified from independent DNA preparations made from individuals. Rainbow trout was distinctly observed 3 specific DNA markers (2. 3, 2.0 and 1.3kb) in bandsharing. Individual fragments generated using the same arbitrary primer, demonstrated that a single primer detected at least three independent genomic polymorphisms in rainbow trout sperm DNA. The RAPD polymorphism generated by this primer may be used as a genetic marker for individual identification The RAPD-PCR technique has been shown to reveal informative polymorphism in many species of fish. The present results demonstrate that RAPD markers are abundant, reproducible and provide a basis for future gene mapping and MAS in these important aquaculture species using RAPD polymorphic markers. It is concluded that RAPD polymorphisms are useful as genetic markers for fish breed differentiation.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.3
• /
• pp.77-90
• /
• 2018

Software Defect Prediction (SDP) is a field of study that identifies defective modules. With insufficient local data, a company can exploit Cross-Project Defect Prediction (CPDP), a way to build a classifier using dataset collected from other companies. Most machine learning algorithms for SDP have used more than one parameter that significantly affects prediction performance depending on different values. The objective of this study is to propose a parameter selection technique to enhance the performance of CPDP. Using a Harmony Search algorithm (HS), our approach tunes parameters of cost-sensitive boosting, a method to tackle class imbalance causing the difficulty of prediction. According to distributional characteristics, parameter ranges and constraint rules between parameters are defined and applied to HS. The proposed approach is compared with three CPDP methods and a Within-Project Defect Prediction (WPDP) method over fifteen target projects. The experimental results indicate that the proposed model outperforms the other CPDP methods in the context of class imbalance. Unlike the previous researches showing high probability of false alarm or low probability of detection, our approach provides acceptable high PD and low PF while providing high overall performance. It also provides similar performance compared with WPDP.

• The Korean Journal of Air & Space Law and Policy
• /
• v.35 no.1
• /
• pp.127-152
• /
• 2020

Despite growing international cooperation for maintenance of international peace and security, wars continue to occur due to conflicted state interests. Continuing conflicts has advanced development of various weapon systems such as global integrated intelligence, surveillance and reconnaissance. However, with a big increase in the number of civilian casualties caused by the weapon systems development, the international community has also advanced diplomatic efforts to minimize deaths of civilian and military personnel. Therefore, it is essential to observe the principle of discrimination between combatants and non-combatants when operating unmanned aerial vehicles (UAVs), better known as drones. Drones have become more capable of distinguishing combatants from non-combatants due to its high-tech prowess. In the operation of drones, any parties involved in combat or the war are responsible for mounting civilian casualties. In addition, it should comply with the principle of proportionality that calls for a balance between results of such action and expected military advantage anticipated from the attack. The rule of proportionality prohibits use of military force which may be expected to cause excessive civilian harm. Drones have been able to track and monitor targets for hours and select the accurate locations of the targets. The aim is to reduce civilian losses and damage to a minimum. Drones meet the standards of Article 51.4 of the Additional Protocol.

• Korean Journal of Remote Sensing
• /
• v.37 no.6_2
• /
• pp.1803-1818
• /
• 2021

As the frequency of drought increases due to climate change, it is very important to have a monitoring system that can accurately determine the situation of widespread drought. However, while ground-based meteorological data has limitations in identifying all the complex droughts in Korea, satellite remote sensing data can be effectively used to identify the spatial characteristics of drought in a wide range of regions and to detect drought. This study attempted to analyze the possibility of using remote sensing data for drought identification in South Korea. In order to monitor various aspects of drought, remote sensing and ground observation data of precipitation and potential evapotranspiration, which are major variables affecting drought, were collected. The evaluation of the applicability of remote sensing data was conducted focusing on the comparison with the observation data. First, to evaluate the applicability and accuracy of remote sensing data, the correlations with observation data were analyzed, and drought indices of various aspects were calculated using precipitation and potential evapotranspiration for meteorological drought monitoring. Then, to evaluate the drought monitoring ability of remote sensing data, the drought reproducibility of the past was confirmed using the drought index. Finally, a high-resolution drought map using remote sensing data was prepared to evaluate the possibility of using remote sensing data for actual drought in South Korea. Through the application of remote sensing data, it was judged that it would be possible to identify and understand various drought conditions occurring in all regions of South Korea, including unmeasured watersheds in the future.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.40 no.2
• /
• pp.91-108
• /
• 2022

In most cases, optical images have been used as training data of DL (Deep Learning) models for object detection, recognition, identification, classification, semantic segmentation, and instance segmentation. However, properties of 3D objects in the real-world could not be fully explored with 2D images. One of the major sources of the 3D geospatial information is DSM (Digital Surface Model). In this matter, characteristic information derived from DSM would be effective to analyze 3D terrain features. Especially, man-made objects such as buildings having geometrically unique shape could be described by geometric elements that are obtained from 3D geospatial data. The background and motivation of this paper were drawn from concept of the intrinsic image that is involved in high-level visual information processing. This paper aims to extract buildings after classifying terrain features by training DL model with DSM-derived information including slope, aspect, and SRI (Shaded Relief Image). The experiments were carried out using DSM and label dataset provided by ISPRS (International Society for Photogrammetry and Remote Sensing) for CNN-based SegNet model. In particular, experiments focus on combining multi-source information to improve training performance and synergistic effect of the DL model. The results demonstrate that buildings were effectively classified and extracted by the proposed approach.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.23 no.3
• /
• pp.27-33
• /
• 2023

Countries around the world are calling for stronger law enforcement to combat the production and distribution of child sexual exploitation images, such as child grooming. Given the scale and importance of this social problem, it requires extensive cooperation between law enforcement, government, industry, and government organizations. In the wake of the Nth Room Case, there have been some amendments to the Enforcement Decree of the Telecommunications Business Act regarding additional telecommunications services provided by precautionary operators in Korea. While Naver and others in Korea use Electronics and Telecommunications Research Institute's own technology to filter illegal images, Microsoft uses its own PhotoDNA technology. Microsoft's PhotoDNA is so good at comparing and identifying illegal images that major global operators such as Twitter are using it to detect and filter images. In order to meet the Korean government's testing standards, Microsoft has conducted more than 16 performance tests on "PhotoDNA for Video 2.0A," which is being applied to the Bing service, in cooperation with the Korea Communications Commission and Telecommunications Technology Association. In this paper, we analyze the cases that did not pass the standards and derive improvement measures related to adding logos. In addition, we propose to use three video datasets for the performance test of filtering against illegal videos.

• Journal of Advanced Navigation Technology
• /
• v.28 no.4
• /
• pp.400-407
• /
• 2024

The role of unmanned aerial vehicles (UAVs) in modern warfare is increasingly significant, making their capacity for autonomous missions essential. Accordingly, autonomous target detection/identification based on captured images is crucial, yet the effectiveness of AI models depends on image sharpness. Therefore, this study describes how to determine the field of view (FOV) of the camera and the flight position of the UAV considering the required spatial resolution. Firstly, the calculation of the size of the acquisition area is discussed in relation to the relative position of the UAV and the FOV of the camera. Through this, this paper first calculates the area that can satisfy the spatial resolution and then calculates the relative position of the UAV and the FOV of the camera that can satisfy it. Furthermore, this paper propose a method for calculating the effective range of the UAV's position that can satisfy the required spatial resolution, centred on the coordinate to be photographed. This is then processed into a tabular format, which can be used for mission planning.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.5
• /
• pp.97-104
• /
• 2024

As indoor air pollution has emerged as a social issue since the COVID-19 pandemic, pollution management in large-scale facilities has been recognized as an important task. For this purpose, this study proposes real-time pollution level detection using sensors and efficient control path setting using Dijkstra algorithm as key technologies. In addition, by introducing outlier determination algorithm and priority algorithm, we propose ways to increase the reliability of the data and enable efficient control work. The outlier determination algorithm describes the process of identifying and processing outliers based on sensor data in an environmental monitoring system. It describes in detail the process of averaging the recent 10 sensor data, calculating the Z-score to detect outliers, and removing and replacing the data determined to be outliers. The priority algorithm describes the process of establishing an efficient control path in consideration of the pollution level of each region. It suggests how to select the most polluted areas first and use them as a starting point to set the control path. In addition, it introduces an iterative process of detecting and responding to the pollution level in real time, which allows the system to be continuously optimized and to respond to environmental pollution. Through this, it is expected to increase the reliability and efficiency of the environmental monitoring system through outlier judgment algorithms and priority algorithms, thereby quickly identifying and responding to pollution situations.