Earticle

비디오 카메라에서 입력된 동영상을 분석하여 교차로에서 교통량을 자동적으로 측정하는 시스템을 설계, 구현하였다. 입력 영상에서 배경과 전경을 분리하기 위하여 3가지 방법을 비교 분석하였으며 이중에서 좋은 결과를 보여주는 Li의 방법을 선택하였다. 전경 영상에서 연결 성분 분석을 이용하여 각각의 블로브들을 분리하였으며 분리된 블로브들은 블로브 추적기를 이용하여 프레임 별로 추적된다. 가장 기본적인 추적기는 블로브의 크기와 위치 정보들을 이용한다. 블로브들 간에 충돌이 있는 경우에는 블로브 안의 컬러 분포를 이용하는 mean shift 알고리즘이 사용되었다. 제안된 시스템은 실제 교차로 동영상을 이용하여 테스트되었으며 휴리스틱을 추가할 경우, 좋은 감지율과 오차율을 보였다.

A vehicle counting system at intersection is designed and implemented using analyzing a video stream from a camera. To separate foreground image from background, we compare three different methods, among which Li's method is chosen. Blobs are extracted from the foreground image using connected component analysis and the blobs are tracked by a blob tracker, frame by frame. The primary tracker use only the size and location of blob in foreground image. If there is a collision between blobs, the mean-shift tracking algorithm based on color distribution of blob is used. The proposed system is tested using real video data at intersection. If some huristics is applied, the system shows a good detection rate and a low error rate.

This paper investigates the non-intrusive speech intelligibility estimation method in noise environments when the bottleneck feature of autoencoder is used as an input to a neural network. The bottleneck featurebased method has the problem of severe performance degradation when the noise environment is changed. In order to overcome this problem, we propose a novel non-intrusive speech intelligibility estimation method that adds the noise environment information along with bottleneck feature to the input of long short-term memory (LSTM) neural network whose output is a short-time objective intelligence (STOI) score that is a standard tool for measuring intrusive speech intelligibility with reference speech signals. From the experiments in various noise environments, the proposed method showed improved performance when the noise environment is same. In particular, the performance was significant improved compared to that of the conventional methods in different environments. Therefore, we can conclude that the method proposed in this paper can be successfully used for estimating non-intrusive speech intelligibility in various noise environments.

Reverse engineering of reliefs aims to turn an existing relief superimposed on an underlying surface into a geometric model which may be applied to a different base surface. Steps in this process include segmenting the relief from the background, and describing it as an offset height field relative to the underlying surface. We have previously considered relief segmentation using a geometric snake. Here, we show how to use this initial segmentation to estimate the background surface lying under the relief, which can be used (i) to refine the segmentation and (ii) to express the relief as an offset field. Our approach fits a B-spline surface patch to the measured background data surrounding the relief, while tension terms ensure this background surface smoothly continues underneath the relief where there are no measured background data points to fit. After making an initial estimate of relief offset height everywhere within the patch, we use a support vector machine to refine the segmentation. Tests demonstrate that this approach can accurately model the background surface where it underlies the relief, providing more accurate segmentation, as well as relief height field estimation. In particular, this approach provides significant improvements for relief concavities with narrow mouths and can segment reliefs with small internal holes.

The airborne radiation monitoring has been used in geophysics for more than forty years and now it also has its important role in emergency monitoring. The aircraft background and the cosmic gamma-rays contribute to the measured gamma spectrum on the aircraft board. This adverse effect should be eliminated before the data processing. The paper describes two semiparametric methods to estimate the full spectrum aircraft background and cosmic gamma-ray contribution from spectra measured at altitudes where terrestrial contribution is negligible. The methods only assume to know possible peak positions in spectra and their full width at half maximum, that can be easily obtained e.g. from terrestrial measurement. The methods were applied to real experimental data acquired on Mi-17 and Bell 412 helicopter boards. The IRIS airborne gamma-ray spectrometer, with 4×4 L NaI(Tl) crystals, produced by Pico Envirotec Inc., Canada, was used on helicopters' boards. To obtain valid estimate of the aircraft background and the cosmic contribution, the measurements over sea and large water areas were carried out. However, the satisfactory results over inland were also achieved comparing with those acquired over large water areas.

It is critically important to identify unexpected loose parts in a nuclear reactor pressure vessel, since they may collide with and cause damage to internal structures. Mass estimation can provide key information regarding the kind as well as the location of loose parts. This study proposes a mass estimation method based on an artificial neural network (ANN), which can overcome several unresolved issues involved in other conventional methods. In the ANN model, input parameters are the discrete cosine transform (DCT) coefficients of the auto-power spectrum density (APSD) of the measured impact acceleration signal. The performance of the proposed method is then evaluated through application to a large-sized plate and a 1/8-scaled mockup of a reactor pressure vessel. The results are compared with those obtained using a conventional method, the frequency ratio (FR) method. It is shown that the proposed method is capable of estimating the impact mass with 30% lower relative error than the FR method, thus improving the estimation performance.

입력영상에 대하여 전처리과정으로 배경을 분리하는 것이 영상처리 및 인식 성능에 중요한 영향을 준다. 본 논문에서는 화재검출을 위한 영상인식 전처리로 활용하는 다양한 배경추정 알고리즘에 대하여 계산량과 배경추정 성능 분석하였다. 비교하는 배경추정알고리즘은 Gaussian Running Average 추정기법, Mixture of Gaussian 모델, 그리고 KDE (kernel density estimate) 알고리즘에 대한 성능을 평가하였다. 입력영상에 대하여 배경영상차로부터 연기를 검출하는데 있어 KDE 알고리즘이 배경추정 성능은 우수한 것을 확인하였다.

The background estimation algorithms had a significant impact on the performance of image processing and recognition. In this paper, background estimation algorithms were analysis of complexity and performance as preprocessing of image recognition. It was evaluated the performance of Gaussian Running Average, Mixture of Gaussian, and KDE algorithm. The simulation results show that KDE algorithm outperforms compared to the other algorithms.

본 논문에서는 카메라로부터 입력된 해상 또는 항만 영상에 대하여 배경추정을 이용한 영상기반의 선박검출과 해당 선박의 AIS 신호를 연동하여 모니터 상에 표출하는 AIS 연동 선박검출시스템을 제안한다. 해상 또는 항만에서 실시간 입력되는 영상에 대하여 선박 검출 실험을 하였다. 시뮬레이션 및 실 환경에서의 실험결과 제안하는 알고리즘이 선박 관제에 효과적인 것을 확인하였다.

In this paper, we propose ship detection system which is co-operated with AIS by using background estimation based on image processing for on the sea or harbor image extracted from camera. We experiment with on the sea or harbor image extracted from real-time input image from camera. By computer simulation and real world test, the proposed system show more effective to ship monitoring.

본 논문에서는 센서 융합과 배경 추정 기법을 이용하여 연속된 영상에서 물체 영역을 검출하는 기법을 제안하였다. IR/CCD각각의 카메라로부터 얻은 입력 영상을 정렬하고 융합하는 과정을 거친 후, 각 화소 단위의 배경 모델을 추정하고 시간이 지남에 따라 이를 갱신함으로써 물체 영역을 효과적으로 검출하는 기법을 제시하고 있다. 실험은 차량을 대상으로 하였고, 카메라가 움직이는 상황과 비교적 복잡한 환경에서도 좋은 결과를 얻을 수 있었다.

본 논문은 초고속영상을 통하여 말 자세를 추정하는 방법에 대한 연구이다. 미디어와 게임 산업의 발전으로 대상의 실제 움직임을 가상의 공간에서 표출하려는 연구가 급속히 증가하고 있다. 하지만 말은 사람에 비해 통제하기가 상대적으로 어려워 데이터 셋을 구축하는 데 제약이 많다. 이로 인해 데이터 셋이 매우 부족하여 연구의 발전 속도가 느리다. 이러한 문제를 해결하기 위해, 본 논문은 새로운 데이터 증강 방법을 제안하고자 한다. 먼저 고속으로 달리는 말을 초고속 카메라로 촬영하여 데이터 셋을 수집하고 전배경 분리와 영상 합성한다. 실험결과 MAE 0.2%~0.8%만큼 정확도가 개선되었고, 기존 Horse-10 데이터셋과 비교할 때 약 MAE 40%만큼 향상된 것을 확인했다.

This paper is a study of horse posture estimation method using Ultra-high speed images. According to the development of media and game industries, researches to express the actual movement of an object in a virtual space are rapidly increasing. But horses are relatively difficult to control compared to people. There are many limitations in building data sets, so the advance of research is slow. To overcome this problem, this paper seeks to propose a novel data augmentation method. First, high-speed horses are photographed with ultra-high-speed cameras to collect data sets, and than background separation and image synthesis are performed. As the result of experiments, it was confirmed that the accuracy was improved by 0.2% to 0.8% of the MAE, and that it was improved by about 40% of the MAE compared to the existing Horse-10 dataset.

It is essential to use specific background concentrations that reliably represent the characteristics of the area under evaluation. In this study, we analyzed the concentrations of heavy metals using the results of the 2022 surveys of the actual state of soil contamination from 17 metropolitan governments. The overall national averages of natural background concentrations were obtained from the results of the 2021-2022 soil monitoring network and compared with that obtained in previous studies. In addition, the natural, anthropogenic, and natural/anthropogenic background concentrations were separately obtained for each of the 17 metropolitan governments. The contamination factor (CF), one of the most commonly used soil pollution indices, was selected and applied to evaluate the soil contamination levels for the 2022 survey results. Variations ranging from 0.3 to 1.4 times were obtained between the national average natural background concentrations from the 2021-2022 network and those from that commonly used in previous studies in Korea. Moreover, the natural and natural/anthropogenic background concentrations of each metropolitan government varied significantly. Compared to the cases using the national average background concentrations, the soil pollution level of the metropolitan governments could be overestimated or underestimated when using the regional average background concentrations.

The purpose of this study was to investigate the effect of background noise on loudness magnitude estimation of Heavy-weight impact noise. Relationship between loudness magnitude estimation and estimation methods about floor impact noise had appeared low in apartment which receive much effect of background noise. Then, to need new estimation method abut effect of background noise. Masking effects by background noise is increased steadily, there is a continuous transition between an audible impact noise and one that is totally masked. Result 1 hat analyze interrelationship of phychoacoustical data and values through Zwicker Parameters, to Investigate that an estimation experiment about Annoyance need.

In this paper, “the Aspect of Harmonization between Being and Non-being of the Buddha-nature” are explained as accurately as possible, and the background of the writing that Wonhyo is presumed to have had in his mind is presented for each topic. First, when the theory of Being of Buddha-nature criticized with the Nirvana Sutra, the theory of Non-being defended with the interpretation on the sutra. While Wonhyo has interpreted that phrases that seem to be different are actually same in the Essentials of the Nirvana Sutra, his method may be used here. Second, when the theory of Being of Buddha-nature criticized with the logical method, the theory of Non-being defended with pointing out that it does not fall under the logical method. While Wonhyo once criticized Xuánzàng’s(玄奘) logical formula, his experience is presumed to be reflected in this topic. Third, when the theory of Non-being criticized with the Treatise of Acclamation of the Sagely Teaching, the theory of Being defended with the interpretation of the sutra. In the Essentials of the Nirvana Sutra, Wonhyo has interpreted sutras with the situation and intention of the author, and it is presumed to be used here. And finally, when the theory of Non-being criticized with assert that all sentient beings will eventually disappear, the theory of Being responded that we should accept it as faith. As Wonhyo emphasized the faith in the Essentials of the Immeasurable Longevity Sutra, his idea may be reflected is this topic.

2019년 7월에 얻어진 200개의 지하수 수질자료를 활용하여 보성지역 지하수의 수질 특성을 평가하고, 배경 수질을 도출하였다. 지하수는 질산성질소 농도를 기준으로 Group 1(NO₃^- < 44.3 mg/L)과 Group 2(NO₃^- ≥ 44.3 mg/L)로 구분하였다. t-test 결과, Group 2는 Group 1에 비해 유의성 있게 수온과 Ca 농도가 높았고, F, As, Pb, Zn 등의 농도는 낮게 나타났다. 질산염 농도가 높은 경우에는 낮은 pH, 알칼리도, NH₄, NO₂, F, As, Fe, Mn, Pb, Zn 등의 농도를 보였으며, EC 및 수온은 높아지는 경향을 보였다. 한편, 배경농도는 BRIDGE 방법으로 설정하였다. 본 연구에서 설정된 NO₃, NH₄, PO₄, As, Fe, Mn 항목들에 대한 배경농도 값들은 기존의 환경부 보고서(MOE and KECO, 2018; 2019)에 제시된 중권역별 배경농도 산정 결과보다는 다소 높게 설정되었다. 본 연구에 이용된 수질자료의 수가 훨씬 많다는 점을 감안하면, 본 연구에서 산출된 배경수질 자료의 신뢰성이 더 높은 것이라 판단된다.

The groundwater quality in Boseong area was characterized and their background concentrations were estimated based on the chemical data obtained from 200 groundwater samples collected during July 2019. Groundwater data were classified into two groups: Group 1 with NO₃^- < 44.3 mg/L and Group 2 with NO₃^- ≥ 44.3 mg/L. Results of t-tests indicate that groundwaters in Group 2 are significantly higher in water temperature and Ca concentration and significantly lower in F, As, Pb and Zn concentrations. It was also revealed that groundwaters Group 2 are closely linked to low pH, alkalinity, NH₄, NO₂, F, As, Fe, Mn, Pb and Zn levels, and high EC and water temperature. Background concentrations were estimated using the the BRIDGE method. The results depicted the higher levels in NO₃, NH₄, PO₄, As, Fe, Mn than the ones estimated by MOE and KECO (2018; 2019), which were prepared for the three catchment units in Boseong. The results of this study are believed to have more reliability because more data were used.

TCP는 Internet 트래픽의 90% 이상을 차지하고 있으며 이러한 TCP의 성능을 이해하는 것이 Internet을 분석하는데 중요하다. TCP의 혼잡 제어 알고리즘 하에서 병목 링크를 공유하는 서로 다른 왕복시간을 갖는 TCP 연결들은 서로 다른 공정성을 갖는다고 알려져 있다. 두 개의 서로 다른 왕복시간을 가지는 TCP의 공정성은 왕복시간의 비율에 반비례하고 동기로 인한 패킷 손실이 일어나면 왕복시간이 긴 TCP는 왕복시간이 짧은 경우보다 더욱 성능이 저하된다. 본 논문은 TCP 동기현상이 백그라운드 트래픽에 의해서도 나타날 수가 있고 이러한 동기현상이 여러 종류의 왕복시간을 가진 TCP 연결이 존재할 때 TCP의 공정성에 어떻게 영향을 주는가를 보여준다. 또한 동기현상을 일으키는 백그라운드 트래픽을 분류할 수 있는 세 가지의 측정방법을 보이며 이 분류를 통해 TCP의 공정성을 예측할 수 있는 방법을 제안한다.

TCP packets occupy over 90% of current Internet traffic thus understanding of TCP throughput is crucial to understand Internet. Under the TCP congestion regime, heterogeneous flows, i.e., flows with different round-trip times (RTTs), that share the same bottleneck link will not attain equal portions of the available bandwidth. In fact, according to the TCP friendly formula, the throughput ratio of two flows is inversely proportional to the ratio of their RTTs. It has also been shown that TCP's unfairness to flows with longer RTTs is accentuated under loss synchronization. In this paper, we show that, injecting bursty background traffic may actually lead to new type of synchronization and result in unfairness to foreground TCP flows with longer RTTs. We propose three different metrics to characterize traffic burstiness and show that these metrics are reliable predictors of TCP unfairness.

본 논문에서는 배경조건을 고려한 Mean-shift 트래킹에서 객체의 크기 추정을 안정적으로 하기 위한 방법을 제안한다. 배경 주색상에 대한 가중치를 낮춘 타겟 모델을 구축한 후 타겟모델과 타겟 후보간의 유사성 판별을 기반으로 커널의 크기를 추정하는 기존 방법에서는 객체의 영역이 작아질수록 더욱 높은 유사도를 보이기 때문에 커널의 shrink 문제가 발생한다. 이 문제를 해결하기 위하여 본 논문에서는 타겟 모델과 타겟 후보간의 유사도를 판별하는 새로운 기준을 제시하고, 이를 기반으로 타겟 후보 중 초기 타겟 모델과 가장 유사한 모습을 가지는 커널 크기를 추정한다. 커널 크기 추정에 있어서 제안 방법의 안정성을 입증하기 위하여 자체 제작한 테스트 영상들과 공개된 테스트 영상들을 사용하여 객체 추적 및 객체 크기 추정 실험을 수행하였고, 기존 방법과의 성능을 비교 분석하였다. 실험 결과 제안방법이 객체 추적 및 객체 크기 추정에 있어서 기존 방법들 보다 낮은 오차율을 보였다.

In this paper, we proposed a robust scale estimation method for mean-shift tracking with background weighted local kernel. The shrinking problem of tracking kernel is frequently happened when the previous scale estimation methods based on color similarity between target and target candidate are applied in the mean-shift tracking with background weighted kernel. In order to solve the kernel shrinking problem, we defined a novel similarity evaluation criterion between target and target candidate based on weight average and found the most similar target candidate among several candidates with different kernel size. In order to validate robustness of the proposed scale estimation method, we compared the proposed method with the previous methods using challenging and self-made test data sequences. Experimental results show the proposed model have good performance to estimate tracked object scale.

본 연구는 구체적인 가구배경 변수들을 사용하여 자녀의 교육성과에 가구 배경과 이웃 등이 미치는 영향을 추정할 때 발생하는 생략변수에 의한 편의 문제를 최소화하기 위해, 형제․자매․남매의 대학 정보를 이용하여 기대수능점수를 계산한 후 이를 가구배경과 이웃효과의 대리변수로 이용하여 그 효과의 크기를 추정하고 하계값을 제시하였다. 추정 결과는 가구배경과 이웃효과가 자녀 교육성과의 20.2% 이상을 설명하는 것으로 나타났으며 평준화 지역만으로 한정할 경우 19.2% 이상을 설명하는 것으로 나타났다. 고졸 형제․자매․남매의 기대수능점수를 계산하여 표본선택 문제를 일부 수정할 경우 23% 이상을, 평준화지역의 경우 20.6% 이상을 설명하는 것으로 나타났다.

In this study, I try to overcome the omitted variable bias that always happens at the regression of educational outcome on observable specific family background and neighborhood variables. To deal with this problem, I estimate the effect of family background and neighborhood on child’s educational outcomes with the college scholastic ability test score of siblings and provide the lower bound of it. The empirical results indicate that the effect of family background and neighborhood explains at least 20.2 percent of the variation of educational outcomes in Korea and 19.2 percent when restricted to the regions under leveling policy. When the high school graduated siblings are included to reduce the sample selection problems, the results show that the size increase to at least 23% in whole region and 20.6% in the region ruled by leveling policy.