• Journal of the Korea Society of Computer and Information
• /
• v.29 no.2
• /
• pp.31-41
• /
• 2024

Deep learning models show excellent performance in tasks such as image classification and object detection in the field of computer vision, and are used in various ways in actual industrial sites. Recently, research on improving robustness has been actively conducted, along with pointing out that this deep learning model is vulnerable to hostile examples. A hostile example is an image in which small noise is added to induce misclassification, and can pose a significant threat when applying a deep learning model to a real environment. In this paper, we tried to confirm the robustness of the edge-learning classification model and the performance of the adversarial example detection model using it for adversarial examples of various algorithms. As a result of robustness experiments, the basic classification model showed about 17% accuracy for the FGSM algorithm, while the edge-learning models maintained accuracy in the 60-70% range, and the basic classification model showed accuracy in the 0-1% range for the PGD/DeepFool/CW algorithm, while the edge-learning models maintained accuracy in 80-90%. As a result of the adversarial example detection experiment, a high detection rate of 91-95% was confirmed for all algorithms of FGSM/PGD/DeepFool/CW. By presenting the possibility of defending against various hostile algorithms through this study, it is expected to improve the safety and reliability of deep learning models in various industries using computer vision.

• Journal of Ecology and Environment
• /
• v.47 no.4
• /
• pp.272-281
• /
• 2023

Ecosystem observatories are burgeoning globally in an endeavour to detect national and global scale trends in the state of biodiversity and ecosystems in an era of rapid environmental change. In this paper we highlight the additional importance of regional scale outcomes of such infrastructure, through an introduction to the Great Western Woodlands TERN (Terrestrial Ecosystem Research Network) SuperSite, and key findings from three gradient plot networks that are part of this infrastructure. The SuperSite was established in 2012 in the 160,000 km² Great Western Woodlands region, in a collaboration involving 12 organisations. This region is globally significant for its largely intact, diverse landscapes, including the world's largest Mediterranean-climate woodlands and highly diverse sandplain shrublands. The dominant woodland eucalypts are fire-sensitive, requiring hundreds of years to regrow after fire. Old-growth woodlands are highly valued by Indigenous and non-Indigenous communities, and managing impacts of climate change and the increasing extent of intense fires are key regional management challenges. Like other TERN SuperSites, the Great Western Woodlands TERN SuperSite includes a core eddy-covariance flux tower measuring exchanges of carbon, water and energy between the vegetation and atmosphere, along with additional environmental and biodiversity monitoring around the tower. The broader SuperSite incorporates three gradient plot networks. Two of these represent aridity gradients, in sandplains and woodlands, informing regional climate adaptation and biodiversity management by characterising biodiversity turnover along spatial climate gradients and acting as sentinels for ecosystem change over time. For example, the sandplains transect has demonstrated extremely high spatial turnover rates in plant species, that challenge traditional approaches to biodiversity conservation. The third gradient plot network represents a 400-year fire-age gradient in Eucalyptus salubris woodlands. It has enabled characterisation of post-fire recovery of vegetation, birds and invertebrates over multi-century timeframes, and provided tools that are directly informing management to reduce stand-replacing fires in eucalypt woodlands. By building regional partnerships and applying globally or nationally consistent methodologies to regional scale questions, ecological observatories have the power not only to detect national and global scale trends in biodiversity and ecosystems, but to directly inform environmental decisions that are critical at regional scales.

• Smart Media Journal
• /
• v.13 no.4
• /
• pp.114-123
• /
• 2024

This paper proposes AR stone tower content, an experiential content based on wearable augmented reality (AR). Although wearable augmented reality is gaining attention, the acceptance of the technology is still focused on specialized applications such as industrial sites. On the other hand, the proposed AR stone tower content is based on the material of 'stone tower' so that general users can relate to it and easily participate in it, and it is organized to utilize space in a moving environment and find and stack stones based on natural hand gestures. The proposed AR stone tower content was implemented in the HoloLens 2 environment and evaluated by general users through a pilot exhibition in a small art museum. The evaluation results showed that the overall satisfaction with the content averaged 3.85, and the content appropriateness for the stone tower material was very high at 4.15. In particular, users were highly satisfied with content comprehension and sound, but somewhat less satisfied with object recognition, body adaptation, and object control. The above user evaluations confirm the resonance and positive response to the material, but also highlight the difficulties of the average user in experiencing and interacting with the wearable AR environment.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.27 no.1
• /
• pp.29-40
• /
• 2024

Recently, the technology of autonomous driving, one of the core of the fourth industrial revolution, is developing, but sensor-based autonomous driving is showing limitations, such as accidents in unexpected situations, To compensate for this, HD-map is being used as a core infrastructure for autonomous driving, and interest in the public and private sectors is increasing, and various studies and technology developments are being conducted to secure the latest and accuracy of HD-map. Currently, NGII will be newly built in urban areas and major roads across the country, including the metropolitan area, where self-driving cars are expected to run, and is working to minimize data error rates through quality verification. Therefore, this study analyzes the spatial relationship of reference objects in the attribute structuring process for rapid and accurate renewal and production of HD-map under construction by NGII, By applying the attribute input automation methodology of the reference object in which spatial relations are established using the library of open source-based PyQGIS, target sites were selected for each road type, such as high-speed national highways, general national highways, and C-ITS demonstration sections. Using the attribute automation tool developed in this study, it took about 2 to 5 minutes for each target location to automatically input the attributes of the spatial relationship reference object, As a result of automation of attribute input for reference objects, attribute input accuracy of 86.4% for high-speed national highways, 79.7% for general national highways, 82.4% for C-ITS, and 82.8% on average were secured.

• Journal of the Korean Geotechnical Society
• /
• v.40 no.4
• /
• pp.33-47
• /
• 2024

Offshore wind structures are subject to long-term repeated horizontal loads from wind, waves, and currents, making it essential to consider these loads in the design of offshore foundations. In this respect, monopiles are large-diameter hollow steel pipes that are relatively simple to construct compared with piles used on onshore sites. They can provide stable support for wind structures and have well-established design codes, leading to their widespread use globally. The behavior of monopiles under lateral static loads is typically assessed using the p-y method proposed by the American Petroleum Institute (API). However, the applicability of p-y curves to large-diameter monopiles exposed to repeated cyclic horizontal loads, such as those experienced in offshore wind applications, must yet be evaluated. Thus, this study evaluated the behavior of monopiles under two-way cyclic horizontal loads in loose silty sand, a representative soil type of the southwestern coast of Korea, using centrifuge model tests. The results demostrated that the behavior of monopiles varied depending on the loading level, number of cycles, and direction of the cyclic loads. Furthermore, the p-y curve method proposed by the API overestimated the behavior of a large-diameter monopile installed in silty sand under two-way cyclic loads.

• Applied Chemistry for Engineering
• /
• v.35 no.4
• /
• pp.335-340
• /
• 2024

In this paper, we synthesized organic and inorganic hybrid materials to introduce antibody functionality to UIO-66 and incorporated them into a surface plasmon resonance (SPR) assay to enhance the sensitivity of detecting small molecules such as oxytocin. A biological marker peptide called oxytocin may help in the diagnosis of heart failure, Alzheimer's disease, and cancer. To detect oxytocin at concentrations as low as a few femtomole (fM), we developed a surface sandwich assay utilizing a pair of oxytocin-specific antibodies for enhancing selectivity and one of metal organic frameworks [e.g., UiO-66-(COOH)₂] possessing high porosity and surface-area as a signal amplifier. Initially, real-time SPR assays were used to confirm that each selected oxytocin-specific antibody binds strongly to oxytocin and to different binding sites on oxytocin. One of these antibodies (e.g., anti-OXT[OTI5G4]) was immobilized on the surface of a thin gold chip. Upon sequential injecting of oxytocin and the other antibody (e.g., anti-OXT[4G11]) conjugated to UiO-66-(COOH)₂ onto the surface to form the surface sandwich complex of anti-OXT[OTI5G4]/oxytocin/UiO-66-(COOH)₂-anti-OXT[4G11]), SPR changes, which varied with oxytocin concentration, were then measured in real time. The results demonstrated that sensitivity was amplified by over a million-fold compared to assays without UiO-66-(COOH)₂, enabling oxytocin detection down to approximately 10 fM.

• Journal of the Korean earth science society
• /
• v.25 no.6
• /
• pp.428-442
• /
• 2004

Dustfall samples were collected by the modified American dust jar (bulk type) at 5 sampling sites in the Nakdong river area from lune 2002 to May 2003. Nineteen chemical species (Al, Ca, Cd, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, Pb, Si, V, Zn, $Cl^-$, $NO_3^-$, $SO_4^{2-}$, and $NH_4^+$) were analyzed via the combination of ICP/AES, AAS, IC and UV. The purposes of this study were to qualitatively evaluate the chemical composition of dustfalls by examining their regional and seasonal distribution patterns. Computation of the enrichment factor showed that well-defined anthropogenic sources, particularly in Pb were found in the order Gamjeondong (industrial area), Wondong, Silla University, Samrangiin and Mulgum. The seasonal mean of soil contribution showed its highest value (16.3%) during the winter with an annual mean of 11.2%. The concentration ratio of [$SO_4^{2-}/NO_3^-$] was found to be highest (5.12) during the winter, while the lowest ratio value (3.30) was seen during the all. fall, Also regional equivalent ratios of [$SO_4^{2-}/NO_3^-$] were found in the order: Silla University (6.78), Gamjeondong (4.98), Mulgum (3.95), Wondong (3.85), and Samrangjin (2.87). Seasonal distribution of water soluble components for total dustfall were found in the order: spring (71.6%), summer (61.2%), fall (49.2%) and winter (48.6%) with a mean ratio of 57.6%. Regional contribution of sea salts of water soluble ions were found in the order: Silla university (34.5%), Gamjeondong (28.3%), Wondong (17.3%), Samrangiin (17.2%) and Mulgum (13.8%), the total mean contribution rate was 22.1%. As for the chemical composition of dustftll on the lower Nakdong river, there is a decreased influence of sea salt and artificial anthropogenic sources and increased influence of soil particle inland. Also, the total amount of deposition on the lower Nakdong river has decreased, with the river's surface serving as a confounding factor in resuspending dusts.

• Journal of the Korean Society for Marine Environment & Energy
• /
• v.17 no.4
• /
• pp.283-296
• /
• 2014

The distribution of heavy metals in the stormwater runoffs from industrial sites around Shihwa Lake that implements the total pollutant load management system (TPLMS) was studied to characterize the temporal changes of metal concentrations and to assess the ecological risk in dissolved and particulate phases of the selected metals. The dissolved Co and Ni concentration demonstrated first flush and tended to decrease with increasing of the duration of rainfall. The intensity of precipitation was found to be the main controlling factor of particulate metals in the stromwater runoffs. The particulate concentration of Pb accounted for 97.2% so the particulate phase was its main form. Other metals followed the sequence: Pb>Cu>Cd>Co>Zn>Ni. The particulate-dissolved partitioning coefficient ($K_d$) indicated that the $K_d$ of Pb were bigger than that of other metals because the metal Pb in the stormwater runoffs is quickly removed into the particulate phase. In a single day rainfall event, total runoff fluxes for total metals as the sum of dissolved and particulate forms through only two sewer outlets were 2.21 kg for Co, 30.5 kg for Ni, 278.3 kg for Cu, 398.3 kg for Zn, 0.39 kg for Cd and 40.0 kg for Pb, respectively. Given the annual rainfall, the number of rain days and the basin area for total pollutant load management system (TPLMS) of Shihwa area, enormous amount of non-point metal pollutants were entered into Lake with any treatment. The dissolved metals (e.g., Ni, Cu and Zn) in the stormwater runoffs exceeded the acute water quality criteria. Additionally, all metals were significantly enriched in the particulate phase and exceeded the PEL criteria of sediment quality guidelines (SQGs). These results indicated that the heavy metals in the stormwater runoffs may pose a very high ecological risk to the coastal environments and ecosystem.

• Journal of Preventive Medicine and Public Health
• /
• v.21 no.1 s.23
• /
• pp.132-151
• /
• 1988

This study was conducted to establish the control program for preventing unfavorable health effects of nitrogen dioxide($NO_2$) exposure in homes by preparing the fundamental data for evaluation of relation-ships between $NO_2$ levels and influencing factors through measurements of indoor-outdoor $NO_2$ levels and personal $NO_2$ exposures for housewives with questionnaire survey on 172 homes in Pusan area from April to June, 1987 $NO_2$ measurements were made by using diffusion tube samplers(Palmes tube $NO_2$ sampler) for one week at 4 sites in homes ; kitchen(KIT), bedroom(BED), living room(LIV), outdoor(OUT) and near the collar of housewives(personal exposure livel, PNO). The details of questionnaire were number of household members(FAM), number of regular smokers (SMOKER), daily number of meals eaten(MEAL), type of housing units(HOUSE), location of house with distance from the heavy traffic roads as walking time(DIST), and of kitchen(KAREA), kind of cooking fuels(FUEL), cooking time of each meal(CTIME), usage of kitchen fan for cooking(FAN), type of heating facilities(HEAT) and so on of subject homes. The Obtained results were as fellows : 1) The mean $NO_2$ level was significantly higher at indoors than outdoors(p<0.01) and the kitchen $NO_2$ level was the highest with $33.7{\pm}13.6ppb$(9.5-81.5ppb). The mean personal exposure level of $NO_2$ for housewives was $20.6{\pm}8.8ppb$(3.1-46.9ppb). 2) The mean indoor $NO_2$ level was significantly higher in the group of household members above 5 than below 4(p<0.05), in detached dwellings than apartments(p<0.001), within 5 minutes of distance than over 5 minutes(p<0.001), in the group of unusing fan(p<0.001), in the group of longer cooking time(p<0.001), and it was in order of coal briquette, gas, electricity and oil by kind of cooking fuels(p<0.05). 3) Variables showing significant correlation(p<0.001) with indoor $NO_2$ level were kitchen $NO_2$ level(r=0.8677), cooking time(r=0.5921), outdoor $NO_2$ level(r=0.5192), personal $NO_2$ exposure level(r=0.4615), usage of kitchen fan(r=0.3573) and location of house(r=-0.2988) 4) As a result of multiple regression analysis, the most significant influencing variable to the kitchen $NO_2$ level was cooking time[KIT=$-0.378{\pm}11.772$(CTIME)+0.298(OUT)+3.102(FAN)], it was kitchen $NO_2$ level to the indoor $NO_2$ level[IND=6.996+0.458(KIT)+0.230(OUT)-1.127(KAREA)], and it was indoor $NO_2$ level to the personal $NO_2$ exposure level[PNO=15.562+0.729(IND)-4.542(DIST)-0.200(KIT)] 5) It was recognized that aritificial ventilation in the kitchen, suppression of unnecessary combustion and replacement of cooking fuel, as much as possible, were effective means for decreasing indoor $NO_2$ levels in homes.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.1
• /
• pp.336-344
• /
• 2017

This study was conducted to identify factors related to the musculoskeletal symptoms of 340 family caregivers who have a patient with brain damage based on self-administered questionnaires. The questionnaires included general characteristics, characteristics of care activities and caregivers' musculoskeletal symptoms. Data were collected from many long-term care hospitals and rehabilitation centers during March 17 to March 21, 2014. The results indicated that complaint rates of work-related musculoskeletal systems of the body sites differed. Factor analysis revealed that neck complaints were related to education (under mid), while shoulder complaints were related to sex (female), age (50-59), education (mid) and duration of care (< 2 years). Arm/elbow complaints were related to age (40-49), education (high) and duration of care (12-24). Complaints regarding the hand/wrist/finger were related to age (50-59) and duration of care (12-24), while those associated with the legs/feet were related to age (50-59) and duration of care (< 6[ED highlight - years? Please specify.]). Back problems were related to education (under mid, mid stage, high) and gait. The complaint rate regarding musculoskeletal symptoms during daily life was not statistically significant. Logistic regression analysis of significant factors related to subjective musculoskeletal symptoms identified ambulation and gait as having the greatest influence and complaint factor among family caregivers. The complaint rate of family caregivers differed among body parts. Being a caregiver for less than one year was found to have a significant impact on pain. Overall, long term family caregivers could be faced with risk factors for musculoskeletal problems, but there are many different factors that affect musculoskeletal symptoms with regards to their activities. Accordingly, comprehensive and systematic prevention plans for family caregivers who have patients with brain damage should be developed.