• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06c
• /
• pp.463-468
• /
• 2010

본 논문은 본 연구팀이 행동패턴 분석을 위하여 개발한 BPP(Behavior Pattern Prediction)알고리즘의 가중치(weight) 속성을 객관적으로 수식화 하는 방법과 가중치와 행동 프로파일을 이용하여 정상/비정상 행동여부를 판단하는 ABA(Abnormal Behavior Analysis) 알고리즘을 제안한다. 가중치는 거주자의 방과 행동 사이의 연관성을 나타내며 가중치가 제한된 범위 내에서 증가 할수록 행동에 대한 관심이 크다. 구축한 사용자 프로파일의 주요 구성 요소로는 행동이 지속된 시간 과 행동 발생 횟수이다. ABA 알고리즘은 가중치와 행동 발생 횟수, 행동 지속시간과의 상관분석 결과를 참조 하였으며, 이산 가중치 데이터를 분석하여 비정상적인 행동을 탐지한다.

• Proceedings of the Korea Institute of Fire Science and Engineering Conference
• /
• 2011.11a
• /
• pp.348-351
• /
• 2011

본 연구는 국내의 급격한 고령화추세 및 등록장애인수 증가에 따른 재해약자시설의 피난안전설계를 위한 기초연구로서 국내 재해약자시설의 화재사례 분석을 통한 문제점 도출과 화재발생 시 인명피해를 최소화 시키기 위해 가장 중요할 것으로 판단되는 피난개시시간 및 초기대응행동에 관한 국외연구사례를 분석하였다. 국내의 경우 피난행동특성에 관한 연구는 전무하며 또한 실제적인 성능설계나 피난계획이 미흡한 실정이므로 향 후 재해약자의 피난행동특성을 고려한 다양한 연구를 통한 DATA확보가 시급할 것으로 판단된다.

• Proceedings of the Korea Contents Association Conference
• /
• 2010.05a
• /
• pp.428-430
• /
• 2010

가상현실에서 실제 사람처럼 행동하는 가상거주자는 스스로 주변의 상황을 판단하고 평가를 내리게 된다. 이러한 상황에 대한 판단은 얼마나 정확하고 다양한 자료가 주어졌느냐에 따라서 달라지게 된다. 본 논문에서는 Spatio-temporal graph(ST graph)를 사용하여 시간과 공간에 대한 데이터를 정의하고, ontology의 개념을 더하여 다양한 상황에 대한 표현이 가능하게 하였다. 이 표현 방법으로 가상거주자는 어떠한 상황을 마주하더라도 주변 환경이나 공간에 대한 데이터를 가지고 분석하여 필요한 행동을 할 수 있게 될 것이다.

• Proceedings of the Korea Institute of Fire Science and Engineering Conference
• /
• 2012.04a
• /
• pp.400-403
• /
• 2012

최근 고시원은 저렴한 방값을 원하는 저소득층의 서민들이 주거하는 공간으로서 불특정 다수가 사용하므로 방화의 위험성을 안고 있다. 본 조사는 이 위험성을 검토하기 위해 SIMULEX 실험을 통하여 고시원 화재시 피난행동을 간접적으로 확인 하고 대비 하는데 도움이 되고자한다. 모형실험 결과 재실자들이 1층 출구 대피시 병목현상을 보였지만 연기의 유동은 위로 상승하기 때문에 3층 재실자들의 빠른 대피가 있다면 인명피해는 없을 것으로 판단된다. 만약 이번 실험 조건이 사람들의 인지능력이 떨어지는 시간 조건을 주고 대피시간이 늦는다면 4, 5층의 재실자는 연기의 빠른 확산으로 인해 많은 인명피해가 발생할 수도 있을 것으로 판단된다.

• Journal of Digital Contents Society
• /
• v.14 no.4
• /
• pp.589-596
• /
• 2013

With the increase of social crime rate, the interest on the intelligent security system is also growing. This paper proposes a detection system of monitoring whether abnormal behavior is being carried in the images captured using CCTV. After detection of an object via subtraction from background image and morpholgy, this system extracts an abnormal behavior by each object's feature information and its trajectory. When an object is loitering for a while in CCTV images, this system considers the loitering as an abnormal behavior and sends the alarm signal to the control center to facilitate prevention in advance. Especially, this research aims at detecting a loitoring act among various abnormal behaviors and also extends to the detection whether an incoming object is identical to one of inactive objects out of image.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.5
• /
• pp.631-637
• /
• 2010

In this paper, We propose a method of motion interpretation of pedestrian for active traffic signal control. We detect pedestrian object in a movie of crosswalk area by using the code book method and acquire contour information. To do this stage fast, we use parallel processing based on CUDA (Compute Unified Device Architecture). And we remove shadow which causes shape distortion of objects. Shadow removed object is judged by using the hilbert scan distance whether to human or noise. If the objects are judged as a human, we analyze pedestrian objects' motion, face area feature, waiting time to decide that they have intetion to across a crosswalk for pdestrians. Traffic signal can be controlled after judgement.

• Journal of Animal Science and Technology
• /
• v.45 no.6
• /
• pp.1089-1096
• /
• 2003

We investigated postpartum behaviors of sows and their litters. The aim of this study was to determine the behavioral recovery of sows and the behavioral development of their piglets postpartum. Five multiparous sows(Landrace${\times}$Yorkshire) and their litters were studied over a period of 5 days postpartum. The litter size varied from 8 to 13 with a mean of 11.2 piglets. The behaviors of sows and their piglets were recorded using five CCD cameras, a multiplexer and a time lapsed VCR. The videotapes were scanned every 2 min to obtain an instantaneous behavioral sample. In the behaviors of sows, ‘Lying ventrally’ increased from 1 h postpartum to 26 h postpartum, then decreased gradually, and stabilized after 80 h postpartum. ‘Lying laterally’ decreased from 1 h postpartum to 26 h postpartum and stabilized after 80 h postpartum. ‘Standing, drinking and feeding’ increased from 1 h postpartum to 23 h postpartum, then decreased gradually, and increased again from 60 h postpartum. ‘Sitting’ increased from 1 h postpartum to 46 h postpartum and then decreased. In the behaviors of piglets, ‘Lying’ increased rapidly from 1h postpartum to 45h postpartum, and then was maintained at an almost constant level. ‘Massaging and Suckling’ decreased rapidly from 1 h postpartum to 36 h postpartum, and then was maintained at an almost constant level. ‘Walking’ increased from 1 h postpartum to 21 h postpartum, then decreased gradually. These results suggest that the behavioral recovery of the sow is almost completed at 80h postpartum and that the behavioral development of the piglet is almost completed at 45h postpartum.

• Journal of KIISE
• /
• v.41 no.12
• /
• pp.1152-1159
• /
• 2014

In this paper, we presents a system to ensure the safety of residents through appropriate action or alarm in case the residents occurs an emergency situation and abnormal activity. We collect and analysis real-time data of living environment of the residents using video and sensor. The existing system have been determined by using only the sensor data it have several problems. Our system attach camera to solve the existing system problem. We use weighted difference image and motion vector. The existing system, it takes about 48 hours to determine that an abnormal activity occurs. However, our system takes less than 1 hour.

• Journal of Korea Game Society
• /
• v.4 no.3
• /
• pp.43-47
• /
• 2004

Fast games such as first person shooting games and action games gives notiming surplus to analyze both states of enemys and my status normally. Behavioral of enemy characters in these games are generally produced by patterns. However It was weekpoints of exposure of patterns when gamers try many times. In this paper, we study the new patterns of enemy characters after analysing patterns of gamers

• Journal of Animal Environmental Science
• /
• v.12 no.1
• /
• pp.13-20
• /
• 2006

This experiment was carried out to determine the effect of breed and sex on behavioral characteristics and economic traits of performance-tested pigs. Totally 32 tested pigs in 4 different breeds were assigned for behavioral observation. Behavior was recorded for 4 hours(each 2 hours, before and after none) of tested pigs in each tested pen. The average daily gain, age at 90 kg and feed efficiency on the basis of the performance data were collected from 8,477 performance-tested pigs in which pure breeds of Duroc, Yorkshire, Landrace and Berkshire at the Korea Swine Testing Station. The results obtained in this study are summarized as follows; The effect of breed was statistically significant for all traits studied. Average daily gain were the highest(P<0.05) in Duroc and ages at 90 kg were the highest(P<0.05) in Berkshire. In feed efficiency of male, Landrace and Yorkshire were the most efficient. The average daily gain and feed efficiency were superior to those of female, however, the age at 90 kg was not different. Overally, females were superior to males in various traits examined. Berkshire breed had more proportion of time spent in ventral tying and sitting whereas the proportion of time spent in walking, drinking and eating was less than that of other breeds. The male pigs showed more time spent in standing and social behavior.