• 한국경영과학회:학술대회논문집
• /
• 대한산업공학회/한국경영과학회 1992년도 춘계공동학술대회 발표논문 및 초록집; 울산대학교, 울산; 01월 02일 May 1992
• /
• pp.436-443
• /
• 1992

Presented in the paper is a systematic approach to constructing a Petri net model of FAS (flexible assembly system). Petri net is widely used in modeling automated manufacturing systems. But, it found to be very difficult for an FA engineer to build a correct model of an FAS with Petri net symbols (ie, place, transition, and token) from the beginning. An automated manufacturing system in general is built from a set of "standard" hardware components. An FAS in particular is usually composed of assembly robots, work tables, conveyor lines, buffer storages, part feeders, etc. In the proposed modeling scheme, each type of standard resources is represented as a standard "module" which is a sub Petri net. Then, the model of a FAS can be conveniently constructed using the predefined modules the same way the FAS itself is built from the standard components. The network representation of a FAS is termed a JR-net (job resource relation net) which is easy to construct. This JR net is then mechanically converted to a formal Petri net (to simulate the behavior of the FAS). The proposed modeling scheme may easily be extended to the modeling of other types of automated manufacturing systems such as FMS and AS/RS.ch as FMS and AS/RS.

• 대한전자공학회논문지TC
• /
• 제43권11호
• /
• pp.35-43
• /
• 2006

본 논문에서는 IETF(Internet Engineering Task Force)에서 진행 중인 NetLMM(Network based Localized Mobility Management) WG의 프로토콜을 이용하여 네트워크 기반의 고속 핸드오버 프로토콜을 제안하였다. NetLMM 프로토콜에서 핸드오버 지연을 개선하기 위해 IEEE 802.21 MIHS(Media Independent Handover Services)를 적용하였으며 Fluid Flow Mobility Model을 이용하여 제안하는 Fast NetLMM의 성능을 분석하였다. 분석 결과 Fast NetLMM 프로토콜은 다른 이동성 관리 프로토콜보다 향상된 성능을 보이는 것을 확인하였다.

• 한국음향학회지
• /
• 제42권5호
• /
• pp.395-401
• /
• 2023

음향 이벤트 검출(Sound Event Detection, SED)은 음향 신호에서 관심 있는 음향의 종류와 발생 구간을 검출하는 기술로, 음향 감시 시스템 및 모니터링 시스템 등 다양한 분야에서 활용되고 있다. 최근 음향 신호 분석에 관한 국제 경연 대회(Detection and Classification of Acoustic Scenes and Events, DCASE) Task 4를 통해 다양한 방법이 소개되고 있다. 본 연구는 다양한 영역에서 성능 향상을 이끌고 있는 Dense Convolutional Networks(DenseNet)을 음향 이벤트 검출에 적용하기 위해 설계 변수에 따른 성능 변화를 비교 및 분석한다. 실험에서는 DenseNet with Bottleneck and Compression(DenseNet-BC)와 순환신경망(Recurrent Neural Network, RNN)의 한 종류인 양방향 게이트 순환 유닛(Bidirectional Gated Recurrent Unit, Bi-GRU)을 결합한 DenseRNN 모델을 설계하고, 평균 교사 모델(Mean Teacher Model)을 통해 모델을 학습한다. DCASE task4의 성능 평가 기준에 따라 이벤트 기반 f-score를 바탕으로 설계 변수에 따른 DenseRNN의 성능 변화를 분석한다. 실험 결과에서 DenseRNN의 복잡도가 높을수록 성능이 향상되지만 일정 수준에 도달하면 유사한 성능을 보임을 확인할 수 있다. 또한, 학습과정에서 중도탈락을 적용하지 않는 경우, 모델이 효과적으로 학습됨을 확인할 수 있다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권5호
• /
• pp.1431-1445
• /
• 2022

We construct a deep neural network model named ECGResNet. This model can diagnosis diseases based on 12-lead ECG data of eight common cardiovascular diseases with a high accuracy. We chose the 16 Blocks of ResNet50 as the main body of the model and added the Squeeze-and-Excitation module to learn the data information between channels adaptively. We modified the first convolutional layer of ResNet50 which has a convolutional kernel of 7 to a superposition of convolutional kernels of 8 and 16 as our feature extraction method. This way allows the model to focus on the overall trend of the ECG signal while also noticing subtle changes. The model further improves the accuracy of cardiovascular and cerebrovascular disease classification by using a fully connected layer that integrates factors such as gender and age. The ECGResNet model adds Dropout layers to both the residual block and SE module of ResNet50, further avoiding the phenomenon of model overfitting. The model was eventually trained using a five-fold cross-validation and Flooding training method, with an accuracy of 95% on the test set and an F1-score of 0.841.We design a new deep neural network, innovate a multi-scale feature extraction method, and apply the SE module to extract features of ECG data.

• 한국인터넷방송통신학회논문지
• /
• 제23권5호
• /
• pp.71-76
• /
• 2023

본 논문에서는 딥러닝을 이용한 영상 분할에서 성능을 향상하기 위해 퍼지 논리를 적용하는 퍼지 딥러닝 모델인 퍼지 U-Net을 제안한다. 퍼지 논리를 이용한 퍼지 모듈을 영상 분할에서 우수한 성능을 보이는 딥러닝 모델인 U-Net에 결합하여 다양한 형태의 퍼지 모듈을 시뮬레이션하였다. 제안된 딥러닝 모델의 퍼지 모듈은 이미지의 특징맵과 해당 분할 결과 사이의 본질적이고 복잡한 규칙을 학습다. 이를 위해 치아 CBCT 데이터에 적용하여 제안된 방법의 우수성을 입증하였다. 시뮬레이션 결과 제안된 퍼지 U-Net에서 더하기 스킵 연결을 사용한 모델의 ADD-RELU 퍼지 모듈 구조의 성능이 시험용 데이터에 대해 0.7928로 가장 우수한 것을 볼 수 있다.

• 수산해양기술연구
• /
• 제29권3호
• /
• pp.200-213
• /
• 1993

A model experiment on a midwater rope trawl net which is used in the North Pacific to catch alaska pollack is carried out in the circulating tank to examine the basic efficiency of the net. The prototype is the net used by M/S Hanil(1, 179GT, 2, 700PS), a Korean trawler. The model net was made according to the Tauti's Similarity Law of Fishing Gear in 1/100 scale by considering the condition of the tank. To measure the basic efficiency of the standard model net, the vertical opening and width between some points marked on the net were measured, and the hydrodynamic resistance were determined. Then the constructive conditions of the net were varied as follows and the factors were measured again to compare the efficiency of those nets with that of the standard net(A-1 type) front weight multiplied 1.5 times: A-2 type. buoyancy and depressing force multiplied 1.7 times: A-3 type. front weight multiplied 1.5 times on A-3 type: A-4 type. depressors rigged at ground rope: B type. cod-end stuffed with cashmylon wad: C type. The results obtained can be summarized as follows: 1. The vertical opening at the center of head rope was steeply decreased with the flow velocity increasing and the vertical opening H(m) can be expressed in H=1.2v super(-1.2)(v : flow velocity in m/sec). The width of the net varied a little when the flow velocity was over 0.4m/sec, and the width of net mouth showed about 37% of the distance between the fore tips of net pendant. The shape of net mouth was almost a circle at 0.2m/sec and then steeply flatted elliptically with the flow velocity increasing and the area of mouth S(m super(2)) can be expressed in S=(1.65-2.3v)$\times$10 super(-2). The hydrodynamic resistance of the net increased almost linearly with the flow velocity increasing and the resistance R(kg) can be expressed in R=3.2$\times$d/l$\times$abv. where d/l denotes the mean of d(diameter of netting twine) and l(length of a leg in a mesh) from wing tip to the end of bag-net except cod-end on the side pannel, and a denotes the strectched circumference of the net at the fore end of a meshed part and b the stretched length of the whole net from wing tip to the end of cod-end. 2. In the condition-varied nets, the vertical opening of head rope showed some increase in every type net except the C type, and the increase showed the greatest in the B type by 30~54%, whereas it showed decrease in the C type by 5~10%. Variation of the area of net mouth showed almost the same tendency as the vertical opening and the increase showed the greatest in the B type by 20%, whereas it showed decrease in the C type by 12%. Hydrodynamic resistance showed some increase in every type compared with the standard net, and the rate of increase indicated 5~10% in the A-2, A-3 and A-4 type, 22% in the B type and 3% in the C type.

• Water Engineering Research
• /
• 제2권3호
• /
• pp.161-169
• /
• 2001

Present study deals with the development of a numerical model for the estimation of net annual recharge by coupling the Galerkin's finite element flow simulationl model with the Gauss-Newton-Marquardt optimization technique. The developed coupled numerical model is applied for estimating net annual recharge for Mahi Right Bank Canal (MRBC) project the norms of Groundwater Resources Estimation committee (1984, 1997) and Indian Agricultural research Institute(1983). It is observed that the estimated net recharge by inverse modeling is closer to the net recharge estimated using the water balance approach. Further it is observed that the computed head distribution from the estimated recharge agree closely with the observed head distribution. The study concludes that the developed model for inverse modeling can be successfully applied to large groundwater system involving regional aquifers where reliable recharge estimation always requires considerable time and financial resources.

• 수산해양기술연구
• /
• 제22권1호
• /
• pp.1-5
• /
• 1986

In order to improve the net-mouth height of dragged gears, the authors devised models of floatingcollars of nylon cloth instead of floats and experimented with 4 types-A type (length 65em, breadth 3em), B type (length 65em, breadth 4em), C type (length 65em, breadth 5em) and D type (length 65 em, breadth 6em) attached respectively to the front edge of square of a model net after preliminary experimentation. These various types of floating collars were experimented in a circulating water channel to evaluate the characteristics of net-mouth height and hydrodynamic resistance and the effect of the length of bridles were also examined. The results obtained were as follows: 1. In case of attaching floats, the model net-mouth height reduced from 80 em to 20 em when current velocity was increased from 0.25m/see to 1m/sec. 2. In case of attaching floating collars, the model net-mouth heights were maintained 70 em, 71 em, 80 em, 78 em in maximum and 55 em, 63 em, 69 em, 73 em in minimum respectively even the current volocity was increased from 0.25 m/see to 1 m/see. 3. The model net-mouth height was reduced to 10 em maximum according to the current velocity and types of floating collars when the bridles were shortened 3~4 mm in length. 4. Hydrodynamic resistance of D type only was increased to 700 g in maximum and those of A, B, C type were reduced to 460 g in maximum at current velocity beyond 0.5 m/ see when bridles were shortened 3-4 mm in length. 5. But the model net-mouth heights became higher in accordance with breadth of floating coliars, B type was the best for this model net in case that hydrodynamic resistance was taken into account.

• 한국컴퓨터그래픽스학회논문지
• /
• 제28권1호
• /
• pp.1-9
• /
• 2022

비소세포폐암(NSCLC)은 전체 폐암 중 85%의 높은 비중을 차지하며 사망률(22.7%)이 다른 암에 비해 현저히 높은 암으로 비소세포폐암 환자의 수술 후 예후에 대한 예측은 매우 중요하다. 본 연구에서는 종양을 관심영역으로 갖는 비소세포폐암 환자의 수술 전 흉부 CT 영상 패치의 종류를 종양 관련 정보에 따라 총 다섯 가지로 다양화하고, 이를 입력데이터로 갖는 사전 학습 된 ResNet 과 EfficientNet CNN 네트워크를 사용하여 단일 모델과 간접 투표 방식을 이용한 앙상블 모델, 그리고 3 개의 입력 채널을 활용한 앙상블 모델에서의 실험 결과 및 성능을 오분류의 사례와 Grad-CAM 시각화를 통해 비교 분석한다. 실험 결과, 종양 주변부 패치를 학습한 ResNet152 단일 모델과 EfficientNet-b7 단일 모델은 각각 87.93%와 81.03%의 정확도를 보였다. 또한 ResNet152 에서 총 3 개의 입력 채널에 각각 영상 패치, 종양 주변부 패치, 형상 집중 종양 내부 패치를 넣어 앙상블 모델을 구성한 경우에는 정확도 87.93%를, EfficientNet-b7 에서 간접 투표 방식으로 영상 패치와 종양 주변부 패치 학습 모델을 앙상블 한 경우에는 정확도 84.48%를 도출하며 안정적인 성능을 보였다.

• 인터넷정보학회논문지
• /
• 제6권6호
• /
• pp.57-69
• /
• 2005

본 논문에서는 워크플로우 프로세스 재발견 문제를 해결하기 위한 적절한 해결책으로 워크플로우 최적 축소 모델을 제안한다. 워크플로우 최적 축소 모델은 워크플로우 최적 축소 넷으로 표현할 수 있다. ICN(Information Control Net) 모델링 기법으로 표현되는 프로세스 모델은 ICN을 구성하고 있는 액티비티 사이의 액티비티 의존성에 따라 적절한 알고리즘의 적용으로 액티비티 의존 넷을 구성할 수 있다. 워크플로우 최적 넷 또한 액티비티 의존 넷의 몇 가지 속성에 대한 알고리즘 적용으로 찾아낼 수 있으며, 찾아낸 워크플로우 최적 넷을 프로세스 재발견 문제를 해결하기 위한 방안으로 제안하며, 프로세스 개선에도 의미를 둘 수 있다.