• Journal of Intelligence and Information Systems
• /
• v.16 no.2
• /
• pp.95-108
• /
• 2010

In this paper, we present a hybrid neural network model for dynamic hand gesture recognition. The model consists of two modules, feature extraction module and pattern classification module. We first propose a modified CNN(convolutional Neural Network) a pattern recognition model for the feature extraction module. Then we introduce a weighted fuzzy min-max(WFMM) neural network for the pattern classification module. The data representation proposed in this research is a spatiotemporal template which is based on the motion information of the target object. To minimize the influence caused by the spatial and temporal variation of the feature points, we extend the receptive field of the CNN model to a three-dimensional structure. We discuss the learning capability of the WFMM neural networks in which the weight concept is added to represent the frequency factor in training pattern set. The model can overcome the performance degradation which may be caused by the hyperbox contraction process of conventional FMM neural networks. From the experimental results of human action recognition and dynamic hand gesture recognition for remote-control electric home appliances, the validity of the proposed models is discussed.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2003.06a
• /
• pp.6-9
• /
• 2003

In this paper, we studied on the difference of speaker adaptation according to the phoneme classification for Korean Speech recognition. In order to study of speech adaptation according to the weight of difference of phoneme as recognition unit, we used SCHMM as recognition system. And Speaker adaptation method used in this paper was MAPE(Maximum A Posteriori Probability Estimation), Linear Spectral Estimation. In order to evaluate the performance of these methods, we used 10 Korean isolated numbers as the experimental data. It is possible for the first and the second methods to be carried out unsupervised learning and used in on-line system. And the first method was shown performance improvement over the second method, and hybrid adaptation showed the better recognition results than those which performed each method. And the result of Speaker adaptation using the variable weight according to the phoneme had better than the result using fixed weight.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.64 no.5
• /
• pp.766-778
• /
• 2015

In this paper, we design a hybrid system for recognition and tracking realized with the aid of polynomial based RBFNNs pattern classifier and particle filter. The RBFNN classifier is built by learning the training data for diverse pose images. The optimized parameters of RBFNN classifier are obtained by Particle Swarm Optimization(PSO). Testing data for pose image is used as a face image obtained under real situation, where the face image is detected by AdaBoost algorithm. In order to improve the recognition performance for a detected image, pose estimation as preprocessing step is carried out before the face recognition step. PCA is used for pose estimation, the pose of detected image is assigned for the built pose by considering the featured difference between the previously built pose image and the newly detected image. The recognition of detected image is performed through polynomial based RBFNN pattern classifier, and if the detected image is equal to target for tracking, the target will be traced by particle filter in real time. Moreover, when tracking is failed by PF, Adaboost algorithm detects facial area again, and the procedures of both the pose estimation and the image recognition are repeated as mentioned above. Finally, experimental results are compared and analyzed by using Honda/UCSD data known as benchmark DB.

• KEPCO Journal on Electric Power and Energy
• /
• v.5 no.3
• /
• pp.157-163
• /
• 2019

K-BEMS System was introduced to reduce peak load and to save total energy of the 120 buildings that KEPCO headquarter and branch offices use. K-BEMS system is composed of PV, battery, and hybrid PCS. In this system, ESS, PV, lighting is used to save building energy based on demand prediction. Currently, neural network technique for short past data is applied to demand prediction, and fixed scheduling method by operator for ESS charging/discharging is used. To enhance this system, KEPCO research institute has carried out this K-BEMS research project for 3 years since January 2016. As the result of this project, we developed new real-time highly reliable building demand prediction technique with error free and optimized automatic ESS charging/discharging technique. Through several field test, we can certify the developed algorithm performance successfully. So we will describe the details in this paper.

• Journal of Digital Convergence
• /
• v.19 no.12
• /
• pp.51-62
• /
• 2021

This study developed the factors capable of systematic/comprehensive evaluation of the task performance in order to strengthen the performance management of the professional career personnel invitation program (PCPIP). To this end, a performance evaluation framework was developed by analyzing existing project evaluation studies based on boundary theory and Kirkpatrick's four-level evaluation model. Afterwords, through two Delphi surveys, evaluation factors that can measure performance in terms of individual and invitation institutions of PCP were derived and validated. With this procedure, five evaluation factors were finally selected: adaptability, connectivity, clarity, compatibility, and expandability. This study has implications suggesting a performance evaluation factors capable of hybrid quantitative/qualitative evaluation for the performance management of PCPIP operated by National Research Foundation of Korea Research since 1994.

• Journal of the Korea Society for Simulation
• /
• v.30 no.1
• /
• pp.75-85
• /
• 2021

In this paper, we present a deep neural network-based prediction model that processes and analyzes the corporate credit and personal credit information of individual business owners as a new method to predict the default rate of individual business more accurately. In modeling research in various fields, feature selection techniques have been actively studied as a method for improving performance, especially in predictive models including many features. In this paper, after statistical verification of macroeconomic indicators (macro variables) and credit information (micro variables), which are input variables used in the default rate prediction model, additionally, through the credit information feature selection method, the final feature set that improves prediction performance was identified. The proposed credit information feature selection method as an iterative & hybrid method that combines the filter-based and wrapper-based method builds submodels, constructs subsets by extracting important variables of the maximum performance submodels, and determines the final feature set through prediction performance analysis of the subset and the subset combined set.

• International Journal of Computer Science & Network Security
• /
• v.22 no.3
• /
• pp.37-44
• /
• 2022

Plug-in Hybrid electric vehicles (PHEV) show great potential to reduce gas emission, improve fuel efficiency and offer more driving range flexibility. Moreover, PHEV help to preserve the eco-system, climate changes and reduce the high demand for fossil fuels. To address this; some basic components and energy resources have been used, such as batteries and proton exchange membrane (PEM) fuel cells (FCs). However, the FC remains unsatisfactory in terms of power density and response. In light of the above, an electric storage system (ESS) seems to be a promising solution to resolve this issue, especially when it comes to the transient phase. In addition to the FC, a storage system made-up of an ultra-battery UB is proposed within this paper. The association of the FC and the UB lead to the so-called Fuel Cell Battery Electric Vehicle (FCBEV). The energy consumption model of a FCBEV has been built considering the power losses of the fuel cell, electric motor, the state of charge (SOC) of the battery, and brakes. To do so, the implementing a reinforcement-learning energy management strategy (EMS) has been carried out and the fuel cell efficiency has been optimized while minimizing the hydrogen fuel consummation per 100km. Within this paper the adopted approach over numerous driving cycles of the FCBEV has shown promising results.

• KNOM Review
• /
• v.24 no.2
• /
• pp.24-34
• /
• 2021

Service providers using edge computing provide a high level of service. As a result, devices store important information in inner storage and have become a target of the latest cyberattacks, which are more difficult to detect. Although experts use a security system such as intrusion detection systems, the existing intrusion systems have low detection accuracy. Therefore, in this paper, we proposed a machine learning model for more accurate intrusion detections of devices in edge computing. The proposed model is a hybrid model that combines a stacked sparse autoencoder (SSAE) and a convolutional neural network (CNN) to extract important feature vectors from the input data using sparsity constraints. To find the optimal model, we compared and analyzed the performance as adjusting the sparsity coefficient of SSAE. As a result, the model showed the highest accuracy as a 96.9% using the sparsity constraints. Therefore, the model showed the highest performance when model trains only important features.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.94-94
• /
• 2023

유역에서의 홍수를 예측하기 위한 다양한 강우-유출 모형들이 개발되어 사용되고 있다. 개념적 강우-유출 모형들은 신뢰성과 적용성이 높아 실무에서 널리 활용되어왔으나, 강우-유출 과정을 단순화하여 고려하므로 유출예측의 정확도에 한계가 있다. 또한 모형의 매개변수에 여러 불확실성이 존재하므로 충분한 양의 관측자료를 사용한 보정 작업이 필요하다. 물리적 강우-유출 모형들은 유출예측 결과가 비교적 물리적으로 정확하다는 장점이 있지만, 높은 계산 비용 및 수치적 불안정성으로 인하여 실무에의 적용이 힘들다. 본 연구에서는 홍수 예측의 정확도와 효율성을 모두 확보할 수 있는 하이브리드 기법을 개발하였다. 본 연구에서 개발한 기법은 물리적 모형인 동역학파 모형과 개념적 모형인 순간단위도 모형, 그리고 딥러닝 모형을 결합하여 사용하는 기법이다. 유역의 조도계수 및 지형을 활용한 동역학파 시뮬레이션을 수행하였으며, 동역학파 시뮬레이션 결과 및 멱함수로 나타내어지는 비선형적 강우-유출 관계를 이용하여 유역의 순간단위도를 유도였다. 또한, 딥러닝 모형인 LSTM 모형을 활용하여 강우손실 매개변수를 추정하였으며, 이를 이용하여 강우손실을 계산한 후 유효강우주상도를 산정하였다. 그리고 유역 출구에서의 홍수수문곡선은 유효강우주상도와 순간단위도를 활용한 회선적분을 통해 예측되었다. 본 연구에서 개발한 기법을 시험유역 및 자연유역에서의 홍수 예측에 적용해보았으며, 예측 결과는 NSE=0.55-0.90, R²=0.67-0.95의 높은 정확도를 보였다. 본 연구에서 유도하는 순간단위도는 한 유역에서 유일하지 않으며, 유효 강우강도의 함수이므로 홍수 예측에 비선형적 강우-유출 관계를 고려할 수 있으며, 수많은 유효 강우강도에 대한 순간단위도들은 멱함수를 이용하여 순간적으로 유도될 수 있다. 또한, 유역의 강우 특성이나 지표면의 토양수분, 식생과 같은 특성을 딥러닝 모형을 통해 고려함으로써 강우 손실 산정의 불확실성을 줄일 수 있다. 또한, 순간단위도 유도를 위한 기초작업인 동역학파 시뮬레이션은 유역의 지형과 조도계수만을 필요로 하므로 미계측 유역에의 적용이 유리하다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.212-212
• /
• 2023

비점오염원관리와 같이 장기적인 유역 관리 계획에서 유역 내 오염원 평가는 정말 중요하다. 유역 내 오염원 평가는 강우 유출에 의한 비점오염 발생원이 어디서 얼마나 발생시키는지에 대한 정량적인 조사가 필요하다. 유역 내의 오염원에 대한 정량적인 조사는 많은 비용과 시간이 필요하다. 이러한 비용과 시간을 줄이기 위해 유역단위 수리 수문 모델을 사용하고 있다. 유역단위 수리수문 모델은 HSPF (Hydrological Simulation Program in Fortran), SWAT (Soil and Water Assessment Tool), L-THIA ACN-WQ(The Long-term Hydrologic Impact Assessment Model with Asymptotic Curve Number Regression Equation and Water Quality model)등 다양한 모델이 사용되고 있다. 하지만 유역 모델을 통한 모의는 다양한 연산 과정을 진행하여 모의까지 많은 시간이 필요하다는 단점이 있다. 이에 따라 데이터 기반 모델링 기법(머신러닝/딥러닝)을 이용한 유출 및 수질 예측 연구가 많이 이루어지고 있다. 단순 머신러닝/딥러닝 기반 모델링 기법은 점(최종유출구)에서의 예측만 가능하여 최적관리 기법 적용 등과 같은 유역관리 방안을 적용하기 힘들다는 문제점이 있다. 따라서 본 연구에서 머신러닝/딥러닝을 통해 일부 수문 프로세스를 대체하고 소유역별 하도추적 기법을 연계하여 유량 및 수질 항목들의 모의가 가능한 하이브리드 모델을 개발하였다. 이는 머신러닝/딥러닝이 유역 모델의 일부 연산 과정을 대체하여 모의시간이 빠르며, 기존 머신러닝/딥러닝 예측 모델에서 평가가 어려웠던 유역 관리 방안 및 최적관리기법 적용 평가에도 활용이 가능할 것으로 판단이 된다.