• 한국산업정보학회논문지
• /
• 제2권1호
• /
• pp.73-87
• /
• 1997

Multilayer neural network using a modified beackpropagation learning algorithm was introduced to achieve automatic identification of different types of aircraft in a variety of 3-D orientations. A 3-D shape of an aircraft can be described by a library of 2-D images corresponding to the projected views of an aircraft. From each 2-D binary aircraft image we extracted 2-D invariant (L, Φ) feature vector to be used for training neural network aircraft classifier. Simulations concerning the neural network classification rate was compared using nearest-neighbor classfier (NNC) which has been widely served as a performance benchmark. And we also introduced reliability measure of the designed neural network classifier.

• 제어로봇시스템학회논문지
• /
• 제5권3호
• /
• pp.338-343
• /
• 1999

In this paper, the fuzzy neural network for the obstacle avoidance, which consists of the straight-line navigation and the barrier elusion navigation, is proposed and examined. For the straight-line navigation, the fuzzy neural network gets two inputs, angle and distance between the line and the mobile robot, and produces one output, steering velocity of the mobile robot. For the barrier elusion navigation, four ultrasonic sensors measure the distance between the barrier and the mobile robot and provide the distance information to the network. Then the network outputs the steering velocity to navigate along the obstacle boundary. Training of the proposed fuzzy neural network is executed in a given environment in real-time. The weights adjusting uses the back-propagation of the gradient of error to be minimized. Computer simulations are carried out to examine the efficiency of the real time learning and the guiding ability of the proposed fuzzy neural network. It has been shown that the mobile robot that employs the proposed fuzzy neural network navigates more safely with and less trembling locus compared with the previous reported efforts.

• 한국정밀공학회지
• /
• 제15권7호
• /
• pp.44-51
• /
• 1998

Two-step neural network is designed for determining electrical discharge machining parameters in low erosion. The first neural network, which is used as a classification network, checks whether the current conditions are appropriate to electrical discharge machining in low tool erosion. If the conditions are appropriate to EDM in low erosion, suitable EDM parameters are generated by the second neural network. Theoretically known EDM conditions are produced and also utilized for training the second neural network. The trained neural network is tested how well suitable EDM machining conditions are generated under unknown machining situations Experimental result shows that the proposed two-step neural network approach could be effectively used for determining EDM parameters in low tool erosion. The results also have a practical contribution to EDM area in that it could be applied for maintaining low tool wear as well as obtaining maximum machining rates simultaneously.

• The Journal of the Acoustical Society of Korea
• /
• 제13권1E호
• /
• pp.64-70
• /
• 1994

예측신경회로망 모델은 패턴 예측에 의한 매우 효과적인 음성인식 모델이다. 그러나, 그러한 모델은 유사한 어휘간에서 변별력이 떨어지는 단점이 있다. 이 논문에서는 그러한 단점을 극복하기 위한 변별력있는 학습 알고리즘을 제안한다. 이 알고리즘은 최소 분류 오차 수식화와 GPD 알고리즘으로부터 유도외면 그에 따라서 인식 오차의 수를 직접 최소화하는 것이 가능하다. 한국어 숫자음에 대한 인식 실험결과, 기존의 알고리즘에서 발생하는 오인식의 30%를 줄일 수 있었다.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1994년도 FIFTH WESTERN PACIFIC REGIONAL ACOUSTICS CONFERENCE SEOUL KOREA
• /
• pp.997-1002
• /
• 1994

Predictive neural network models are powerful speech recognition models based on a nonlinear pattern prediction. Those models can effectively normalize the temporal and spatial variability of speech signals. But those models suffer from poor discrimination between acoustically similar words. In this paper, we propose a discriminative training algorithm for predictive neural network models based on a generalized probabilistic descent (GPD) algorithm and minimum classification error formulation (MCEF). The Evaluation of our training algorithm on ten Korean digits shows its effectiveness by 40% reduction of recognition error.

• ETRI Journal
• /
• 제45권4호
• /
• pp.557-569
• /
• 2023

Traditional deterministic channel modeling is accurate in prediction, but due to its complexity, improving computational efficiency remains a challenge. In an alternative approach, we investigated a multilayer artificial neural network (ANN) to predict large-scale and small-scale channel characteristics in metro tunnels. Simulated high-precision training datasets were obtained by combining measurement campaign with a ray tracing (RT) method in a metro tunnel. Performance on the training data was used to determine the number of hidden layers and neurons of the multilayer ANN. The proposed multilayer ANN performed efficiently (10 s for training; 0.19 ms for prediction), and accurately, with better approximation of the RT data than the single-layer ANN. The root mean square errors (RMSE) of path loss (2.82 dB), root mean square delay spread (0.61 ns), azimuth angle spread (3.06°), and elevation angle spread (1.22°) were impressive. These results demonstrate the superior computing efficiency and model complexity of ANNs.

• Computers and Concrete
• /
• 제25권5호
• /
• pp.401-409
• /
• 2020

Artificial neural networks are used as a useful tool in distinct fields of civil engineering these days. In order to control long-term quality of Self-Compacting Semi-Lightweight Concrete (SCSLC), the 90 days compressive strength is considered as a key issue in this paper. In fact, combined artificial neural networks are used to predict the compressive strength of SCSLC at 28 and 90 days. These networks are able to re-establish non-linear and complex relationships straightforwardly. In this study, two types of neural networks, including Radial Basis and Multilayer Perceptron, were used. Four groups of concrete mix designs also were made with two water to cement ratios (W/C) of 0.35 and 0.4, as well as 10% of cement weight was replaced with silica fume in half of the mixes, and different amounts of superplasticizer were used. With the help of rheology test and compressive strength results at 7 and 14 days as inputs, the neural networks were used to estimate the 28 and 90 days compressive strengths of above-mentioned mixes. It was necessary to add the 14 days compressive strength in the input layer to gain acceptable results for 90 days compressive strength. Then proper neural networks were prepared for each mix, following which four existing networks were combined, and the combinatorial neural network model properly predicted the compressive strength of different mix designs.

• 정보처리학회논문지B
• /
• 제14B권4호
• /
• pp.255-262
• /
• 2007

본 연구는 칼라 이미지에서 인물의 얼굴 영역을 검출하는 개선된 색상 기반 방식을 제안한다. 제안 방법은 RGB, $YC_bC_r$, YIQ의 세 가지 색상 모델을 조합, 각각 휘도와 색도 성분 조합 히스토그램을 구축하고 구축된 색상 조합 히스토그램을 역전파방식의 신경망에 입력한 후 학습단계의 반본 과정에 Levenberg-Marquadt 알고리즘을 적용한다. 제안 방법은 신경망 학습과정에 Levenberg-Marquadt 알고리즘을 적용하여 얼굴 검출에 가장 많이 사용되는 방법 중 하나인 역전파 신경망이 지역 최소값에 봉착하는 문제점을 해결함으로써 검출 오류율을 낮추는데 기여한다. 또한 색상 조합 히스토그램을 사용한 새로운 색상 조합 기반의 얼굴 영역 검출 방법은 빛의 영향에 강건하도록 휘도 성분을 분리하고 색도 성분을 강조하여 단일 색상 히스토그램보다 신경망에 더 신뢰성 있는 값을 입력함으로써 단일 색상 공간을 사용했을 때보다 높은 얼굴 검출율을 보인다. 실험 결과는 제안 방식이 얼굴 영역 검출 개선에 효과적이며 빛의 변화에 강건함을 보여준다.

• 한국음향학회지
• /
• 제41권2호
• /
• pp.235-241
• /
• 2022

수동 음향 관측을 통해 수집된 방대한 양의 데이터에서 해양포유류의 소리를 탐지하고 식별하기 위해 합성곱 신경망(Convolutional Neural Network, CNN)을 활용한 연구가 많이 수행되고 있다. 본 연구는 2017년 8월부터 2018년 8월까지 동시베리아 해에서 수집된 수중음향 스펙트럼 이미지를 기반으로 CNN을 활용하여 턱수염물범 소리의 분류 자동화 가능성을 확인해 보았다. 학습 데이터로서 다른 소음이 거의 포함되지 않은 뚜렷한 턱수염물범 소리를 사용하였을 때, 암기로 인한 과적합이 발생하였다. 일부 데이터를 소음이 포함된 데이터로 교체하여 학습시켜 수집된 전체 데이터로 평가한 결과 정확도(0.9743), 정밀도(0.9783), 재현율(0.9520)으로 모델이 이전보다 일반화되어 과적합이 방지되는 것을 확인하였다. 본 연구를 통해 물범신호 분류는 학습 데이터에 소음이 포함되었을 때 성능이 증가하는 것으로 나타났다.

• Computers and Concrete
• /
• 제24권6호
• /
• pp.555-560
• /
• 2019

During the application of shotcrete, a part of the concrete bounces back after hitting to the surface, the reinforcement or previously sprayed concrete. This rebound material is definitely not added to the mixture and considered as waste. In this study, a deep neural network model was developed to predict the rebound material during shotcrete application. The factors affecting rebound and the datasets of these parameters were obtained from previous experiments. The Long Short-Term Memory (LSTM) architecture of the proposed deep neural network model was used in accordance with this data set. In the development of the proposed four-tier prediction model, the dataset was divided into 90% training and 10% test. The deep neural network was modeled with 11 dependents 1 independent data by determining the most appropriate hyper parameter values for prediction. Accuracy and error performance in success performance of LSTM model were evaluated over MSE and RMSE. A success of 93.2% was achieved at the end of training of the model and a success of 85.6% in the test. There was a difference of 7.6% between training and test. In the following stage, it is aimed to increase the success rate of the model by increasing the number of data in the data set with synthetic and experimental data. In addition, it is thought that prediction of the amount of rebound during dry-mix shotcrete application will provide economic gain as well as contributing to environmental protection.