• Journal of the Korea Society of Computer and Information
• /
• v.29 no.1
• /
• pp.69-75
• /
• 2024

This study focuses on enhancing the understanding of the form and characteristics of traditional towers, or 'pagodas,' by utilizing WebXR technology to enable users to explore 3D models and experience them in virtual reality on the web. Traditional towers in Korea pose challenges for direct on-site verification due to their size, making it difficult to examine the structure and features of each level. To address these issues, this research aims to provide users with a WebXR service that allows them to remotely explore and analyze towers without geographical or temporal constraints. The research methodology involves utilizing WebAR to offer a web-based service where users can directly view the original form of the tower's 3D model using smart devices both online and on-site. However, outdoor conditions may affect performance, and to address this, a tower-outline detection and matching technique was employed. Consequently, we propose a remote support service for traditional towers, allowing users to remotely access information and features of various towers nationwide on the web. Meanwhile, on-site visits can involve experiencing augmented reality representations of towers using smart devices.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.83-88
• /
• 2024

In this research, we propose a Semi-Supervised learning based railroad surface defect detection method. The Resnet50 model, pretrained on ImageNet, was employed for the training. Data without labels are randomly selected, and then labeled to train the ResNet50 model. The trained model is used to predict the results of the remaining unlabeled training data. The predicted values exceeding a certain threshold are selected, sorted in descending order, and added to the training data. Pseudo-labeling is performed based on the class with the highest probability during this process. An experiment was conducted to assess the overall class classification performance based on the initial number of labeled data. The results showed an accuracy of 98% at best with less than 10% labeled training data compared to the overall training data.

• Journal of Food Hygiene and Safety
• /
• v.28 no.1
• /
• pp.69-74
• /
• 2013

This study has been performed for 150 days from February 1 - June 31, 2012 aiming at analyzing biologically hazardous factors in order to develop HACCP system for the vinegared pickle radishes. A process chart was prepared as shown on Fig. 1 by referring to manufacturing process of manufacturer of general vinegared pickle radishes regarding process of raw agricultural products of vinegared pickle radishes, used water, warehousing of additives and packing material, storage, careful selection, washing, peeling off, cutting, sorting out, stuffing (filling), internal packing, metal detection, external packing, storage and consignment (delivery). As a result of measuring Coliform group, Staphylococcus aureus, Salmonella spp., Bacillus cereus, Listeria Monocytogenes, E. coli O157:H7, Clostridium perfringens, Yeast and Mold before and after washing raw radishes, Bacillus cereus was $5.00{\times}10$ CFU/g before washing but it was not detected after washing and Yeast and Mold was $3.80{\times}10^2$ CFU/g before washing but it was reduced to 10 CFU/g after washing and other pathogenic bacteria was not detected. As a result of testing microorganism variation depending on pH (2-5) of seasoning fluid (condiment), pH 3-4 was determined as pH of seasoning fluid as all the bacteria was not detected in pH3-4. As a result of testing air-borne bacteria (number of general bacteria, colon bacillus, fungus) depending on each workplace, number of microorganism of internal packing room, seasoning fluid processing room, washing room and storage room was detected to be 10 CFU/Plate, 2 CFU/Plate, 60 CFU/Plate and 20 CFU/Plate, respectively. As a result of testing palm condition of workers, as number of general bacteria and colon bacillus was represented to be high as 346 $CFU/Cm^2$ and 23 $CFU/Cm^2$, respectively, an education and training for individual sanitation control was considered to be required. As a result of inspecting surface pollution level of manufacturing facility and devices, colon bacillus was not detected in all the specimen but general bacteria was most dominantly detected in PP Packing machine and Siuping machine (PE Bulk) as $4.2{\times}10^3CFU/Cm^2$, $2.6{\times}10^3CFU/Cm^2$, respectively. As a result of analyzing above hazardous factors, processing process of seasoning fluid where pathogenic bacteria may be prevented, reduced or removed is required to be controlled by CCP-B (Biological) and threshold level (critical control point) was set at pH 3-4. Therefore, it is considered that thorough HACCP control plan including control criteria (point) of seasoning fluid processing process, countermeasures in case of its deviation, its verification method, education/training and record control would be required.

• Korea journal of population studies
• /
• v.10 no.2
• /
• pp.17-45
• /
• 1987

This study contributes to understanding women's labor market behavior by focusing on a particular set of labor force transitions - labor force withdrawal and entry during the period surrounding the first birth of a child. In particular, this study provides a dynamic analyses, using longitudinal data and event history analysis, to conceptualize labor force behaviors in a straightforward way. The main research question addresses which factors increase or decrease the hazard rates of leaving and entering the labor market. This study used piecewise Gompertz model, following the guide of the non-parametric analysis on the hazard rates, which allowed relatively detailed description on the distribution of timing of leave and entry to the labor market as parameters of interest. The results show that preferences and structural variables, as well as economic considerations, are very important factors to explain the labor market behavior of women in the period surrounding childbirth.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.7
• /
• pp.519-527
• /
• 2003

Speech is much influenced by the existence of outliers which are introduced by such an unexpected happenings as additive background noise, change of speaker's utterance pattern and voice detection errors. These kinds of outliers may result in severe degradation of speaker recognition performance. In this paper, we proposed the GMM based on robust principal component analysis (RPCA-GMM) using M-estimation to solve the problems of both ouliers and high dimensionality of training feature vectors in speaker identification. Firstly, a new feature vector with reduced dimension is obtained by robust PCA obtained from M-estimation. The robust PCA transforms the original dimensional feature vector onto the reduced dimensional linear subspace that is spanned by the leading eigenvectors of the covariance matrix of feature vector. Secondly, the GMM with diagonal covariance matrix is obtained from these transformed feature vectors. We peformed speaker identification experiments to show the effectiveness of the proposed method. We compared the proposed method (RPCA-GMM) with transformed feature vectors to the PCA and the conventional GMM with diagonal matrix. Whenever the portion of outliers increases by every 2%, the proposed method maintains almost same speaker identification rate with 0.03% of little degradation, while the conventional GMM and the PCA shows much degradation of that by 0.65% and 0.55%, respectively This means that our method is more robust to the existence of outlier.

• Journal of the Korean Society of Radiology
• /
• v.14 no.3
• /
• pp.211-220
• /
• 2020

High-quality computed tomography (CT) images enable precise lesion detection and accurate diagnosis. A lot of studies have been performed to improve CT image quality while reducing radiation dose. Recently, deep learning-based techniques for improving CT image quality have been developed and show superior performance compared to conventional techniques. In this study, a super-resolution convolutional neural network (SRCNN) model was used to improve the spatial resolution of CT images, and image quality according to the hyperparameters, which determine the performance of the SRCNN model, was evaluated in order to verify the effect of hyperparameters on the SRCNN model. Profile, structural similarity (SSIM), peak signal-to-noise ratio (PSNR), and full-width at half-maximum (FWHM) were measured to evaluate the performance of the SRCNN model. The results showed that the performance of the SRCNN model was improved with an increase of the numbers of epochs and training sets, and the learning rate needed to be optimized for obtaining acceptable image quality. Therefore, the SRCNN model with optimal hyperparameters is able to improve CT image quality.

• Proceedings of the Korean Society of Organic Agriculture Conference
• /
• 2009.12a
• /
• pp.307-307
• /
• 2009

우리나라에서 농가단위 경축순환농업 모델에 대한 조사와 농장내 순환구조에서 양분수지를 조사한 사례도는 미미한 실정이다. 이에 경축연계 자원순환 유기농업농가의 실천사례 발굴하고, 실천모델별 축산형태 및 경종형태를 조사하여 경종부분의 양분순환과 양분수지를 조사하고자 하였다. 발굴된 농가단위 경축순환농업 사례는 3가지 형태로 모두 한우를 11~21두 범위에서 사육하는 동시에 $15,510m^2$의 밭농사를 수행하는 농가, $8,019m^2$의 밭농사와 $8,250m^2$ 논농사를 동시에 수행하는 농가, $26,400m^2$의 논농사만 수행하는 농가들이었다. 각 모델에서 배합사료는 모두 외부에서 구입하고 있었으며 조사료의 자급율은 25 ~ 100%인 것으로 조사되었다. 특히, 한우 20두와 논농사 $26,400m^2$를 경영하는 농가에서 조사료(볏짚)의 100%를 자급하였고 동시에 한우사육과정에서 발생한 자급퇴비를 전량 논농사에 사용하여 유기농 벼농사를 유지하고 있었다. 밭농사를 함께 수행하는 농가에서는 자급퇴비 외에 외부로부터 유박비료 및 발효유기질비료를 구입하여 양분을 충당하고 있었다. 각 농가의 토양이화학성을 분석한 결과 pH는 5.49~7.61, 유기물 함량은 1.37~1.40%, 유효인산 함량 253.32~329.63 mg/kg, 칼륨 0.24~0.3, 칼슘 3.97~10.1, 마그네슘 0.89~1.77 $Cmol^+$/kg, CEC는 7.92~11 $Cmol^+$/kg 이었므며 토양내 잔류농약은 검출되지 않았다. 농가별로 발생한 우분퇴비의 성분 분석결과는 전질소 0.68 ~ 0.73%, 전인산 0.68 ~ 0.86%, 칼륨 0.70~0.78% 수준이었다. 각 사례농가별 투입한 실제시비량, 토양분석결과와 사용된 자재의 성분 분석결과를 이용한 시비처방법에 따른 시비량 및 표준시비법에 따른 시비량을 산출하여 3요소의 양분수지를 계산하였다. 이와 함께 유기질비료의 무기화율을 감안한 시비량도 산출하였다. 양분수지를 분석한 결과 3농가 모두 실제시비량은 3요소 모두 초과 되는 것으로 나타났으며, 특히 인산과 칼륨이 2배정도 초과되는 경향을 보였다. 그러나 투입된 자급퇴비 및 유기질비료의 무기화율을 감안한 시비량으로 환산할 경우에는 질소성분이 3농가 모두 부족한 것으로 산출되었으며, 인산과 가리 성분은 충분하거나 초과되는 것으로 계산되었다. 농장내 축산경영을 통해 발생하는 자급퇴비만을 이용할 경우에 경종부문의 양분수지를 산출한 결과를 보면 실제시비량 기준으로 질소는 56~186%, 인산은 90~346%, 칼륨은41~221%로 양분수지가 달라졌으며, 무기화를 감안한 기준으로는 질소는 16~53%, 인산은 52~197%, 칼륨은 41~221%로 양분수지가 달라졌다. 소규모 유기농가단위 경축연계 자원순환 모델 연구를 통해 유기경종농업이 유지될 수 있으나, 3요소별 양분수지의 불균형이 발생할 수 있는 것으로 조사되었으며, 유기질비료의 특성상 무기화율을 감안한 시비량을 적용할 경우에는 질소 성분의 부족과 동시에 인산, 칼륨 성분의 과다가 예측되었다. 따라서 이러한 성분의 불균형을 조정할 시비체계 연구가 필요한 것으로 판단되었다. 본 연구는 농촌진흥청의 "유기가축사양 및 유기 순환체계모델" 연구사업의 세부과제로 수행한 1년차 결과입니다.

• The KIPS Transactions:PartD
• /
• v.18D no.4
• /
• pp.261-274
• /
• 2011

In recent competitive business environment each enterprise has to be agile and flexible. For these purposes run-time monitoring ofservices provided by an enterprise and early decision making through this becomes core competition of the enterprise. In addition, in order to process various innumerable events which are generated on enterprise systems techniques which make filtering of meaningful data are needed. However, the existing study related with this is nothing but discovering of service faults by monitoring depending upon API of BPEL engine or middleware, or is nothing but processing of simple events based on low-level events. Accordingly, there would be limitations to provide useful business information. In this paper, through situation detection an extended complex event model is presented, which is possible to provide more valuable and useful business information. Concretely, first of all an event processing architecture in an enterprise system is proposed, and event meta-model which is suitable to the proposed architecture is going to be defined. Based on the defined meta-model, It is presented that syntax and semantics of constructs in our event processing language including various and progressive event operators, complex event pattern, key, etc. In addition, an event context mechanism is proposed to analyze more delicate events. Finally, through application studies application possibility of this study would be shown and merits of this event model would be present through comparison with other event model.

• Korean Journal of Remote Sensing
• /
• v.23 no.4
• /
• pp.297-309
• /
• 2007

The first Korean geostationary weather satellite, Communications, Oceanography and Meteorology Satellite (COMS) will be launched in 2008. The ground station for COMS needs to perform geometric correction to improve accuracy of satellite image data and to broadcast geometrically corrected images to users within 30 minutes after image acquisition. For such a requirement, we developed automated and fast geometric correction techniques. For this, we generated control points automatically by matching images against coastline data and by applying a robust estimation called RANSAC. We used GSHHS (Global Self-consistent Hierarchical High-resolution Shoreline) shoreline database to construct 211 landmark chips. We detected clouds within the images and applied matching to cloud-free sub images. When matching visible channels, we selected sub images located in day-time. We tested the algorithm with GOES-9 images. Control points were generated by matching channel 1 and channel 2 images of GOES against the 211 landmark chips. The RANSAC correctly removed outliers from being selected as control points. The accuracy of sensor models established using the automated control points were in the range of $1{\sim}2$ pixels. Geometric correction was performed and the performance was visually inspected by projecting coastline onto the geometrically corrected images. The total processing time for matching, RANSAC and geometric correction was around 4 minutes.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.1099-1110
• /
• 2023

The recognition system of autonomous driving and robot navigation performs vision work such as object recognition, tracking, and lane detection after multi-sensor fusion to improve performance. Currently, research on a deep learning model based on the fusion of a camera and a lidar sensor is being actively conducted. However, deep learning models are vulnerable to adversarial attacks through modulation of input data. Attacks on the existing multi-sensor-based autonomous driving recognition system are focused on inducing obstacle detection by lowering the confidence score of the object recognition model.However, there is a limitation that an attack is possible only in the target model. In the case of attacks on the sensor fusion stage, errors in vision work after fusion can be cascaded, and this risk needs to be considered. In addition, an attack on LIDAR's point cloud data, which is difficult to judge visually, makes it difficult to determine whether it is an attack. In this study, image scaling-based camera-lidar We propose an attack method that reduces the accuracy of LCCNet, a fusion model (camera-LiDAR calibration model). The proposed method is to perform a scaling attack on the point of the input lidar. As a result of conducting an attack performance experiment by size with a scaling algorithm, an average of more than 77% of fusion errors were caused.