• 한국산업정보학회논문지
• /
• 제16권4호
• /
• pp.9-18
• /
• 2011

본 논문에서는 모델 기반으로 추정한 사람의 시선 방향을 이용하여 실내 환경에서 발생 할 수 있는 사람의 행동을 인식하는 방법을 제안한다. 제안하는 방법은 크게 두 단계로 구성된다. 첫째, 행동 인식을 위한 사전 정보를 얻는 단계로 사람의 머리 영역을 검출하고 시선 방향을 추정한다. 사람의 머리 영역은 색상 정보와 모양 정보를 이용하여 검출하고, 시선 방향은 머리와 얼굴의 관계를 표현한 베이지안 네트워크 모델을 이용하여 추정한다. 둘째, 이벤트와 사람의 행동을 나타내는 시나리오를 인식하는 단계이다. 이벤트는 사람의 상태 변화로 인식하고, 시나리오는 이벤트들의 조합과 제약 사항을 이용하여 규칙 기반으로 인식한다. 본 논문에서는 시선방향과 연관이 있는 4 가지의 시나리오를 정의하여 실험 한다. 실험을 통해 시선 방향 추정의 성능과 시선 방향이 고려된 상황에서의 행동 인식 성능을 보인다.

• 대한전자공학회논문지SP
• /
• 제47권6호
• /
• pp.39-47
• /
• 2010

영상 기반의 모션 캡처에 대한 연구는 인체의 특징 영역 검출, 정확한 자세 추정 및 실시간 성능 등의 문제를 풀기 위해 많은 연구가 진행되고 있다. 특히, 인체의 많은 관절 정보를 복원하기 위해 다양한 방법이 제안되고 있다. 본 논문에서는 수치적인 역운동학 방법의 단점을 개선한 실시간 모션 캡처 방법을 제안한다. 기존의 수치적인 역운동학 방법은 많은 반복 연산이 필요하며, 국부최소치 문제가 발생할 수 있다. 본 논문에서는 이러한 문제를 해결하기 위해 기존의 수치적인 역운동학 해법과 UKF를 결합하여 중간관절을 복원하는 방법을 제안한다. 수치적인 역운동학의 해와 UKF를 결합함으로써, 중간 관절 추정 시 최적값에 보다 안정적이고 빠른 수렴이 가능하다. 모션 캡처를 위해 먼저, 배경 차분과 피부색 검출 방법을 이용하여 인체의 특징 영역을 추출한다. 다수의 카메라로부터 추출된 2차원 인체 영역 정보로부터 3차원 정보를 복원하고, UKF와 결합된 수치적인 역운동학 해법을 통해 동작자의 중간 관절 정보를 추정한다. 수치적인 역운동학의 해는 UKF의 상태 추정 시 안정적인 방향을 제시하고, UKF는 다수의 샘플을 기반으로 최적 상태를 찾음으로써, 전역해에 보다 빠르게 수렴한다.

• 한국환경과학회지
• /
• 제25권4호
• /
• pp.525-534
• /
• 2016

The biological wastewater treatment plant, which uses microbial community to remove organic matter and nutrients in wastewater, is known as its nonlinear behavior and uncertainty to operate. Therefore, operation of the biological wastewater treatment process much depends on observation and knowledge of operators. The manual inspection of human operators is essential to manage the process properly, however, it is impossible to detect a fault promptly so that the process can be exposed to improper condition not securing safe effluent quality. Among various process faults, equipment malfunction is critical to maintain normal operational state. To detect equipment faults automatically, the dynamic time warping was tested using on-line oxidation-reduction potential (ORP) and dissolved oxygen (DO) profiles in a sequencing batch reactor (SBR), which is a type of wastewater treatment process. After one cycle profiles of ORP and DO were measured and stored, they were warped to the template profiles which were prepared already and the distance result, accumulated distance (D) values were calculated. If the D values were increased significantly, some kinds of faults could be detected and an alarm could be sent to the operator. By this way, it seems to be possible to make an early detecting of process faults.

• Archives of Pharmacal Research
• /
• 제19권6호
• /
• pp.469-474
• /
• 1996

A rapid and sensitive method has been developed to detect hepatitis B virus DNA (HBV) by nested polymerase chain reaction (PCR) technique with primers specific for the surface and core regions in capillary thermal cycler within 80 min. The lower limit for detection by present PCR method is $10^{-5}$ pg of recombinant HBV DNA which is equivalent to that determined by one round of PCR amplification and Southern blot hybridization analysis. When boiled HBV positive serum was serially diluted 10-fold, HBV DNA was successfully determined in $1{\mu}l-10^{-3}$ of serum. HBV DNA was detected by present method in 69 clinical samples including HBsAg positives and negatives by enzyme-linked immunosorbent assay (ELISA). When serum samples were amplified by nested PCR using surface and core region primers, HBV DNAs were detected in 37 of 69 samples (53.6%) and 18 of 69 samples (26.1%), respectively. These results can inform the infectious state of HBsAg positive pateints. A simple and rapid nested PCR protocol by using boiled serum as DNA template has been described for the clinical utility to determine HBV DNA in human serum.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2005년도 추계종합학술대회
• /
• pp.705-709
• /
• 2005

In this paper, we introduce fuzzy algorithm similar to human's way of thinking and designed collision detection system of vehicles. First, before the model vehicles design, we did simulation collision detection using PID and Fuzzy Controller. As a result, P.O that is Percent Overshoot when make use of PID controller happened from smallest 32% to 45%. But, In case of using fuzzy controller they produced about 10% in 7% in case use 25 rule. We designed model vehicles that introduce Auto Guided Vehicle(AGV) with confirmed result in simulation. We set Polaroid 6500 sensor on the front of model automobile because distinguish existence automobile to the head. And we composed motor drive part to run vehicles and 80C196KC processor for control movement of vehicles influenced on distance data of the front vehicles that receive from supersonic waves sensor. In case of using Fuzzy controller, last value percent error happened about maximum 15% in smallest 5%, and we confirmed that distance with front vehicles kept when state hold time is about maximum 16 seconds in smallest 10 seconds.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 춘계학술대회
• /
• pp.141-143
• /
• 2021

The Activity patterns of animal species are difficult to access and the behavior of freely moving individuals can not be assessed by direct observation. As it has become large challenge to understand the activity pattern of animals such as dogs, and cats etc. One approach for monitoring these behaviors is the continuous collection of data by human observers. Therefore, in this study we assess the activity patterns of dog using the wearable sensors data such as accelerometer and gyroscope. A wearable, sensor -based system is suitable for such ends, and it will be able to monitor the dogs in real-time. The basic purpose of this study was to develop a system that can detect the activities based on the accelerometer and gyroscope signals. Therefore, we purpose a method which is based on the data collected from 10 dogs, including different nine breeds of different sizes and ages, and both genders. We applied six different state-of-the-art classifiers such as Random forests (RF), Support vector machine (SVM), Gradient boosting machine (GBM), XGBoost, k-nearest neighbors (KNN), and Decision tree classifier, respectively. The Random Forest showed a good classification result. We achieved an accuracy 86.73% while the detecting the activity.

• 대한인간공학회지
• /
• 제30권6호
• /
• pp.705-713
• /
• 2011

Objective: The aim of this study is to compare results of emotion recognition by several algorithms which classify three different emotional states(happiness, neutral, and surprise) using physiological features. Background: Recent emotion recognition studies have tried to detect human emotion by using physiological signals. It is important for emotion recognition to apply on human-computer interaction system for emotion detection. Method: 217 students participated in this experiment. While three kinds of emotional stimuli were presented to participants, ANS responses(EDA, SKT, ECG, RESP, and PPG) as physiological signals were measured in twice first one for 60 seconds as the baseline and 60 to 90 seconds during emotional states. The obtained signals from the session of the baseline and of the emotional states were equally analyzed for 30 seconds. Participants rated their own feelings to emotional stimuli on emotional assessment scale after presentation of emotional stimuli. The emotion classification was analyzed by Linear Discriminant Analysis(LDA, SPSS 15.0), Support Vector Machine (SVM), and Multilayer perceptron(MLP) using difference value which subtracts baseline from emotional state. Results: The emotional stimuli had 96% validity and 5.8 point efficiency on average. There were significant differences of ANS responses among three emotions by statistical analysis. The result of LDA showed that an accuracy of classification in three different emotions was 83.4%. And an accuracy of three emotions classification by SVM was 75.5% and 55.6% by MLP. Conclusion: This study confirmed that the three emotions can be better classified by LDA using various physiological features than SVM and MLP. Further study may need to get this result to get more stability and reliability, as comparing with the accuracy of emotions classification by using other algorithms. Application: This could help get better chances to recognize various human emotions by using physiological signals as well as be applied on human-computer interaction system for recognizing human emotions.

• 로봇학회논문지
• /
• 제12권3호
• /
• pp.313-321
• /
• 2017

One of the most frequently performed tasks in human-robot interaction (HRI), intelligent vehicles, and security systems is face related applications such as face recognition, facial expression recognition, driver state monitoring, and gaze estimation. In these applications, accurate head pose estimation is an important issue. However, conventional methods have been lacking in accuracy, robustness or processing speed in practical use. In this paper, we propose a novel method for estimating head pose with a monocular camera. The proposed algorithm is based on a deep neural network for multi-task learning using a small grayscale image. This network jointly detects multi-view faces and estimates head pose in hard environmental conditions such as illumination change and large pose change. The proposed framework quantitatively and qualitatively outperforms the state-of-the-art method with an average head pose mean error of less than $4.5^{\circ}$ in real-time.

• International Journal of Computer Science & Network Security
• /
• 제23권4호
• /
• pp.48-54
• /
• 2023

Nowadays, a blind man finds it very difficult to cross the roads. They should be very vigilant with every step they take. To resolve this problem, Convolutional Neural Networks(CNN) is a best method to analyse the data and automate the model without intervention of human being. In this work, a traffic signal recognition system is designed using CNN for the visually impaired. To provide a safe walking environment, a voice message is given according to light state and timer state at that instance. The developed model consists of two phases, in the first phase the CNN model is trained to classify different images captured from traffic signals. Common Objects in Context (COCO) labelled dataset is used, which includes images of different classes like traffic lights, bicycles, cars etc. The traffic light object will be detected using this labelled dataset with help of object detection model. The CNN model detects the color of the traffic light and timer displayed on the traffic image. In the second phase, from the detected color of the light and timer value a text message is generated and sent to the text-to-speech conversion model to make voice guidance for the blind person. The developed traffic light recognition model recognizes traffic light color and countdown timer displayed on the signal for safe signal crossing. The countdown timer displayed on the signal was not considered in existing models which is very useful. The proposed model has given accurate results in different scenarios when compared to other models.

• 한국의학물리학회:학술대회논문집
• /
• 한국의학물리학회 2002년도 Proceedings
• /
• pp.356-358
• /
• 2002

MRI, particularly diffusion weighted imaging (DWI), plays vital roles in detection of the acute brain infarction$\^$1-4/ and others metabolic changes of biological tissues. In general, every molecule in biological tissues may diffuse and move randomly in three-dimensional space. However, in clinical diagnosis, only 2D-DWI is used. The authors have developed a new method for rapid three-dimensional DWI (3D-DWI). In this method, by refocusing of the magnetized spin with the applied gradient field, direction of which is opposite to phase encoding field. Magnetized spin of $^1$H is kept under the SSFP (steady state free precession)$\^$5-6/. Under SSFP, in addition of FID, spin echo and stimulated echo are also generated, so the acquired signal is increased. The signal intensity is increased depending on flip angle (FA) of magnetized spin. This phenomenon is confirmed by human brain and phantom studies. The performance of this method is quantitatively analyzed by using both of conventional spin echo DWI and 3D-DWI. From experimental results, three dimensional diffusion weighted images are obtained correctly for liquid phantoms (water, acetone and oil), diffusion coefficient is enhanced in each image. Therefore, this method will provide useful information for clinical diagnosis.