• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권1호
• /
• pp.169-189
• /
• 2015

Detection of anomalous events from video streams is a challenging problem in many video surveillance applications. One such application that has received significant attention from the computer vision community is traffic video surveillance. In this paper, a Lossy Count based Sequential Temporal Pattern mining approach (LC-STP) is proposed for detecting spatio-temporal abnormal events (such as a traffic violation at junction) from sequences of video streams. The proposed approach relies mainly on spatial abstractions of each object, mining frequent temporal patterns in a sequence of video frames to form a regular temporal pattern. In order to detect each object in every frame, the input video is first pre-processed by applying Gaussian Mixture Models. After the detection of foreground objects, the tracking is carried out using block motion estimation by the three-step search method. The primitive events of the object are represented by assigning spatial and temporal symbols corresponding to their location and time information. These primitive events are analyzed to form a temporal pattern in a sequence of video frames, representing temporal relation between various object's primitive events. This is repeated for each window of sequences, and the support for temporal sequence is obtained based on LC-STP to discover regular patterns of normal events. Events deviating from these patterns are identified as anomalies. Unlike the traditional frequent item set mining methods, the proposed method generates maximal frequent patterns without candidate generation. Furthermore, experimental results show that the proposed method performs well and can detect video anomalies in real traffic video data.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 V
• /
• pp.2879-2882
• /
• 2003

Obstructive Sleep Apnea(OSA) is a representative symptom of sleep disorder which is caused by airway obstruction. OSA is usually diagnosed through the laboratory based Polysomnography(PSG) which is uncomfortable and expensive. In this paper, the detection method for OSA events, using ECG, has been developed. The proposed method uses the ECG data sets provided from Physionet. The features for OSA events detection are the average and standard deviation of 1 minute R-R interval, power spectrum of R-R interval and S-pulse amplitude from data sets. These features are applied to the input of Neural Network. To evaluate the method, we used the another ECG data sets. And we achieved sensitivity of 89.66％, specificity of 95.25％. So, we can know that the features proposed in this paper are important to detect OSA.

• 한국통계학회:학술대회논문집
• /
• 한국통계학회 2004년도 학술발표논문집
• /
• pp.243-248
• /
• 2004

The timings of two successive events of interest may not be measurable, instead it may be right censored or interval censored; this data structure is called doubly censored data. In the study of HIV, two such events are the infection with HIV and the onset of AIDS. These data have been analyzed by authors under the assumption that infection time and induction time are independent. This paper investigates the regression problem when two events arc modeled to allow the presence of a possible relation between two events as well as a subject-specific effect. We derive the estimation procedure based on Goetghebeur and Ryan's (2000) piecewise exponential model and Gauss-Hermite integration is applied in the EM algorithm. Simulation studies are performed to investigate the small-sample properties and the method is applied to a set of doubly censored data from an AIDS cohort study.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2004년도 학술대회 논문집 정보 및 제어부문
• /
• pp.325-327
• /
• 2004

ECG limb lead II signal widely used to diagnosis heart diseases and it is essential to detect ECG events (onsets, offsets and peaks of the QRS complex P wave and T wave) and extract them from ECG signal for heart diseases diagnoses. However, it is very difficult to develop standardized feature extraction formulas since ECG signals are varying on patients and disease types. In this paper, simple feature extraction method from normal and abnormal types of ECG signals is proposed. As a signal features, heart rate, PR interval, QRS interval, QT interval, interval between S wave and baseline, and T wave types are extracted. To show the validity of proposed method, Right Bundle Branch Block (RBBB), Left Bundle Branch Block (LBBB), Sinus Bradycardia, and Sinus Tachycardia data from MIT-BIH arrhythmia database are used for feature extraction and the extraction results showed higher extraction capability compare to conventional formula based extraction method.

• 한국콘텐츠학회논문지
• /
• 제11권1호
• /
• pp.56-64
• /
• 2011

스트림데이터는 무한하고 연속적인 특성을 지니고 있기 때문에 전체 데이터를 기반으로 빈발 항목 집합을 탐사하는 것은 어렵다. 이 때문에 데이터의 특성과 사용자의 특성을 반영한 특수한 데이터마이닝 방법이 필요하다. 이 논문에서는 사용자가 최근에 발생한 데이터에 더 많은 관심이 있다는 특성을 반영하여 빈발 항목을 탐사하는 FIMWB 방법을 제안한다. FIMWB는 과거 데이터의 발생 시점과 현재 시점과의 시간 간격에 따라 가변적인 가중치를 배치에 부여하여 최신 데이터에 더 많은 관심과 중요성을 반영한다. FP-Digraph는 FIMWB를 통해 탐사된 빈발 항목으로 그래프를 구성하여 빈발 항목 집합을 탐사한다. 실험 결과로 FIMWB 방법이 불필요한 항목의 생성을 감소시키고 트리기반(FP-Tree)의 빈발 항목 집합 탐사에 비해 제안하는 FP-Digraph 방법이 스트림 데이터 환경에 더 적합함을 알 수 있다.

• 한국습지학회지
• /
• 제18권1호
• /
• pp.76-83
• /
• 2016

최근 연속적인 태풍에 의한 일련의 극한 호우 사상으로 홍수가 발생하였고, 이로 인해 인명과 막대한 재산피해가 발생하였다. 본 연구에서는 연속 호우 사상으로 인해 발생한 극한홍수를 거대홍수라고 정의하고, 일정 시간 간격으로 극한 호우 사상이 연속적으로 발생 될 수 있음을 가정하여 가상의 거대홍수 시나리오를 구성하였다. 최소 무강우 시간 결정(Inter Event Time Definition, IETD)방법을 사용하여 연속적인 강우의 시간 간격을 결정하였으며, IETD에 의해 산정된 시간 간격 안에서 호우 사상을 연속적으로 발생시켜 평창강 유역을 대상으로 거대홍수를 모의하였다. 즉, (1) 기록된 극한 호우 사상의 연속적인 발생 (2) 기왕 자료를 기반으로 빈도해석에 의해 산정된 설계 호우 사상의 연속적인 발생을 가정하여 거대홍수를 모의하였다. 연속 호우 사상으로 인한 거대홍수는 단일 호우 사상으로 인한 일반 홍수에 비해 6~17%의 홍수량이 증가하는 것으로 나타났다. 앞의 호우 사상으로 인한 홍수량에 비해 뒤에 오는 호우로 인한 홍수량의 증가는 많지 않지만, 연속적인 호우는 두 번의 홍수피해를 가져오므로 가상의 거대홍수로 인한 홍수 피해는 매우 클 것으로 판단된다. 따라서 본 연구와 같이 가상의 강우 시나리오를 통해 예상하지 못한 연속적인 홍수 재해와 같은 비상 상황에 대비할 방안을 마련할 필요가 있을 것으로 사료된다.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제2권1호
• /
• pp.27-34
• /
• 2013

배경제거는 동영상의 내용을 자동으로 분석하기 위한 매우 중요한 기술의 하나로 움직이는 객체를 검출하고 추적하기 위한 핵심 기술이다. 본 논문에서는 배경 모델과 함께 배경 영상을 제공하는 새로운 샘플링 기반의 배경제거 알고리즘을 제안한다. 제안된 방법에서는 움직임이 빠른 객체와 느린 객체를 동시에 처리하기 위해 다중 구간 샘플링 기법을 이용하여 배경 모델을 생성한다. 이러한 다중 구간 배경 모델들로부터 최선의 배경 모델을 만들기 위해 "신뢰도"를 사용한 것이 본 논문의 특징이다. 배경 제거 분야에서 다양한 모델을 병합하여 하나의 모델을 만들기 위해 신뢰도를 정의하여 사용한 경우는 현재까지 보고되지 않았다. 실험을 통해 제안된 방법이 다양한 속도의 객체가 존재하고 시간에 따른 그림자의 이동과 같은 환경 변화가 있는 응용에서도 안정적인 결과를 나타내는 것을 알 수 있었다.

• 한국물환경학회지
• /
• 제22권5호
• /
• pp.856-864
• /
• 2006

For estimating the minimum discharge to maintain a river, low flow analysis is required and long term runoff records are needed for the analysis. However, runoff data should be estimated to run a hydrologic model for ungaged river basin. For the reason, parameter estimation is crucial to simulate rainfall-runoff events for those basins using Tank model. In this study, only runoff data recorded for dry season are used for parameter estimation, which is different to other methods based on runoff data recorded for wet and dry seasons. The Harmony Search algorithm is used to determine the optimum parameters for Tank model. The coefficient of determination ($R^2$) is served as the objective function in the Harmony Search. In cases that recorded data are insufficient, the recording interval is changed and Empirical CDF is adopted to analyze the estimated parameters. The suggested method is applied to Yongdam dam, Soyanggang dam, Chungju dam and Seomjingang dam basins. As results, the higher $R^2s$ are obtained when the shorter recording interval, the better recorded data quality, and the more rainfall events recorded along with certain rainfall amount is. Moreover, when the total rainfall is higher than the certain amount, $R^2$ is high. Considering the facts found from this study for the low flow analysis, it is possible to estimate the parameters for Tank model properly with the desired confidence level.

• 지질동맥경화학회지
• /
• 제7권2호
• /
• pp.110-121
• /
• 2018

Objective: We developed a new equation for predicting coronary heart disease (CHD) risk in Korean diabetic patients using a hospital-based cohort and compared it with a UK Prospective Diabetes Study (UKPDS) risk engine. Methods: By considering patients with type 2 diabetes aged ${\geq}30years$ visiting the diabetic center in Boramae hospital in 2006, we developed a multivariable equation for predicting CHD events using the Cox proportional hazard model. Those with CHD were excluded. The predictability of CHD events over 6 years was evaluated using area under the receiver operating characteristic (AUROC) curves, which were compared using the DeLong test. Results: A total of 732 participants (304 males and 428 females; mean age, $60{\pm}10years$; mean duration of diabetes, $10{\pm}7years$) were followed up for 76 months (range, 1-99 month). During the study period, 48 patients (6.6%) experienced CHD events. The AUROC of the proposed equation for predicting 6-year CHD events was 0.721 (95% confidence interval [CI], 0.641-0.800), which is significantly larger than that of the UKPDS risk engine (0.578; 95% CI, 0.482-0.675; p from DeLong test=0.001). Among the subjects with <5% of risk based on the proposed equation, 30.6% (121 out of 396) were classified as ${\geq}10%$ of risk based on the UKPDS risk engine, and their event rate was only 3.3% over 6 years. Conclusion: The UKPDS risk engine overestimated CHD risk in type 2 diabetic patients in this cohort, and the proposed equation has superior predictability for CHD risk compared to the UKPDS risk engine.

• 조명전기설비학회논문지
• /
• 제27권4호
• /
• pp.16-22
• /
• 2013

This paper presents a method to analyze the voltage sag data obtained from monitoring systems. In order to establish effective countermeasures against voltage sag problems, an assessment of the system performance with respect to voltage sags is needed. Generally, the average annual sag frequency can be estimated by using the recorded voltage sag events for several years. However, the simple average value can not give the information about the errors of estimation. Such an average estimation is not useful for establishing effective solutions for voltage sag problems. Therefore, this paper proposes an effective method based on the Interval Estimation method. The estimation of voltage sag frequency is performed by using the average frequency and Poisson probability model. The proposed method can give the expected annual sag frequency and upper one-sided bound frequency.