http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Crowd Activity Recognition using Optical Flow Orientation Distribution
( Jinpyung Kim ),( Gyujin Jang ),( Gyujin Kim ),( Moon-hyun Kim ) 한국인터넷정보학회 2015 KSII Transactions on Internet and Information Syst Vol.9 No.8
In the field of computer vision, visual surveillance systems have recently become an important research topic. Growth in this area is being driven by both the increase in the availability of inexpensive computing devices and image sensors as well as the general inefficiency of manual surveillance and monitoring. In particular, the ultimate goal for many visual surveillance systems is to provide automatic activity recognition for events at a given site. A higher level of understanding of these activities requires certain lower-level computer vision tasks to be performed. So in this paper, we propose an intelligent activity recognition model that uses a structure learning method and a classification method. The structure learning method is provided as a K2-learning algorithm that generates Bayesian networks of causal relationships between sensors for a given activity. The statistical characteristics of the sensor values and the topological characteristics of the generated graphs are learned for each activity, and then a neural network is designed to classify the current activity according to the features extracted from the multiple sensor values that have been collected. Finally, the proposed method is implemented and tested by using PETS2013 benchmark data.
Kim, Chanhoon,Song, Gyujin,Luo, Langli,Cheong, Jun Young,Cho, Su-Ho,Kwon, Dohyung,Choi, Sungho,Jung, Ji-Won,Wang, Chong-Min,Kim, Il-Doo,Park, Soojin American Chemical Society 2018 ACS NANO Vol.12 No.8
<P>Nanowires (NWs) synthesized <I>via</I> chemical vapor deposition (CVD) have demonstrated significant improvement in lithium storage performance along with their outstanding accommodation of large volume changes during the charge/discharge process. Nevertheless, NW electrodes have been confined to the research level due to the lack of scalability and severe side reactions by their high surface area. Here, we present nanoporous Ge nanofibers (NPGeNFs) having moderate nanoporosity <I>via</I> a combination of simple electrospinning and a low-energetic zincothermic reduction reaction. In contrast with the CVD-assisted NW growth, our method provides high tunability of macro/microscopic morphologies such as a porosity, length, and diameter of the nanoscale 1D structures. Significantly, the customized NPGeNFs showed a highly suppressed volume expansion of less than 15% (for electrodes) after full lithation and excellent durability with high lithium storage performance over 500 cycles. Our approach offers effective 1D nanostructuring with highly customized geometries and can be extended to other applications including optoelectronics, catalysis, and energy conversion.</P> [FIG OMISSION]</BR>
Estimation of Crowd Density in Public Areas Based on Neural Network
( Gyujin Kim ),( Taeki An ),( Moonhyun Kim ) 한국인터넷정보학회 2012 KSII Transactions on Internet and Information Syst Vol.6 No.9
There are nowadays strong demands for intelligent surveillance systems, which can infer or understand more complex behavior. The application of crowd density estimation methods could lead to a better understanding of crowd behavior, improved design of the built environment, and increased pedestrian safety. In this paper, we propose a new crowd density estimation method, which aims at estimating not only a moving crowd, but also a stationary crowd, using images captured from surveillance cameras situated in various public locations. The crowd density of the moving people is measured, based on the moving area during a specified time period. The moving area is defined as the area where the magnitude of the accumulated optical flow exceeds a predefined threshold. In contrast, the stationary crowd density is estimated from the coarseness of textures, under the assumption that each person can be regarded as a textural unit. A multilayer neural network is designed, to classify crowd density levels into 5 classes. Finally, the proposed method is experimented with PETS 2009 and the platform of Gangnam subway station image sequences.
김규진(Gyujin Kim),민경덕(Kyoungdoug Min) 한국자동차공학회 2015 한국자동차공학회 학술대회 및 전시회 Vol.2015 No.11
A two dimensional flamelet model, which uses two mixture fraction variables, was introduced to represent the non-premixed combustion of multiple injections. However, the model’s computational time drastically increased due to the expansion of the solution domain. Thus, a modified 2-D flamelet model was introduced to reduce the computational time of the two dimensional flamelet model. The objective of this study is to extend the modified 2-D flamelet model to three or more injection strategies without increasing the computational costs. A few multiple injection strategies (pilot-pilot-main/pilot-pilot-main-post) were applied to a multi-cylinder engine. Two operating conditions (1500/4, 1500/6, [rpm/bar]) were tested, and the simulation results of the pressure using the modified 2-D flamelet model were compared with the experimental data. Lastly, the combustion and emission characteristics of each injection strategy were determined, and the model demonstrated a good agreement with the experimental results.