• 대한수의학회지
• /
• 제53권4호
• /
• pp.207-210
• /
• 2013

Fifteen dry dentate and complete mandible samples from the White Rasquera goat breed were studied for symmetry. Thirty-one landmarks were digitally located on the images of the lateral and medial aspects of each hemimandible. Distances between these landmarks allowed the evaluation of the whole hemimandible and also the neural mandible. In the studied samples, the mandible was rather symmetrical, especially in the medial neural part, and in general, there was no side dominance. Only the diastema differed significantly between the sides, and this was related to the rostral part (incisive arch). The incisive region was the least symmetrical region of the caprine mandible, indicating a modular structure more conservative for the neural part. If unsigned asymmetry is interpreted as a measure of developmental stability, then the studied breed presented a marked ability to develop in good fitness despite the harsh environment. The measurements presented here can also be used as a reference for researchers designing experimental studies, especially on mandibular catch-up growth, and as an aid for zooarchaeologists comparing results from dead animals with those from living goat populations.

• 한국정보처리학회논문지
• /
• 제4권12호
• /
• pp.3063-3068
• /
• 1997

일반적인 역전파(Back Propagation)의 알고리즘을 다층 다항식 고차 뉴럴네트워크에 적용하여 위치 변환 패턴에 있어 정확한 인식을 할 수 있도록 네트워크의 구조를 개선했다. 본 논문의 목적은 다층 다항식 고차 뉴럴네트워크를 이용하여 여러 가지 패턴 인식이 가능한 이유를 자세히 논한다. 실제 예로는 일정한 문자 인식의 예제로 변형된 영문자 T-C 패턴을 가지고 실험했으며, 네트워크의 일반성(Generalization)을 측정하기 위해서는 거울반사 대칭(Mirror Symmetry)문제를 시뮬레이션 했다. 그 결과 종래의 모델보다 기술적인 우수성을 확인 할 수가 있었다. 본 연구가 제안한 방식에 의한 위치 변환된 T-C 패턴에 대하여서는 90%의 인식율을 얻을 수 있었으며, 일반성의 실험에서 거울반사 대칭(Mirror Symmetry)에 대한 인식율은 70％를 얻었다. 이 실험결과는 종래의 모델에서는 구하기 어려운 인식율이며 기존 연구와 비교한 결과 본 제안 방식의 기술적 우위성을 확연히 판단 할 수 있다.

• Nuclear Engineering and Technology
• /
• 제52권12호
• /
• pp.2709-2716
• /
• 2020

Compositions of large nuclear cores (e.g. boiling water reactors) are highly heterogeneous in terms of fuel composition, control rod insertions and flow regimes. For this reason, they usually lack high order of symmetry (e.g. 1/4, 1/8) making it difficult to estimate their neutronic parameters for large spaces of possible loading patterns. A detailed hyperparameter optimization technique (a combination of manual and Gaussian process search) is used to train and optimize deep neural networks for the prediction of three neutronic parameters for the Ringhals-1 BWR unit: power peaking factors (PPF), control rod bank level, and cycle length. Simulation data is generated based on half-symmetry using PARCS core simulator by shuffling a total of 196 assemblies. The results demonstrate a promising performance by the deep networks as acceptable mean absolute error values are found for the global maximum PPF (~0.2) and for the radially and axially averaged PPF (~0.05). The mean difference between targets and predictions for the control rod level is about 5% insertion depth. Lastly, cycle length labels are predicted with 82% accuracy. The results also demonstrate that 10,000 samples are adequate to capture about 80% of the high-dimensional space, with minor improvements found for larger number of samples. The promising findings of this work prove the ability of deep neural networks to resolve high dimensionality issues of large cores in the nuclear area.

• Journal of the Korean Physical Society
• /
• 제73권9호
• /
• pp.1385-1392
• /
• 2018

A network with coupled biological neurons provides various forms of collective synchronous dynamics. Such phase-locking dynamics states resemble eigenvectors in a linear coupling system in that the forms are determined by the symmetry of the coupling strengths. However, the states behave as attractors in a nonlinear dynamics system. We here study the collective synchronous dynamics in a neural system by using a novel theory. We exhibit how the period and the stability of individual phase-locking dynamics states are determined by the characteristics of synaptic couplings. We find that, contrary to common sense, the firing rate of a synchronized state decreases with increasing synaptic coupling strength.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2000년도 추계학술발표논문집 (상)
• /
• pp.243-246
• /
• 2000

얼굴 영역 검출은 주어진 영상에서 얼굴의 유.무, 개수 및 위치를 검출하는 것으로 본 논문은 영상에서 얼굴을 검출하는 방법으로 신경망(Neural Network)을 적용하였다. 검출률의 향상 및 오검출률의 감소, 계산량을 최대한 줄이기 위해 후보 영역의 최적화와 얼굴의 대칭성(symmetry of face)을 이용한 좌우 평균 명암도 비교방법, 평균 얼굴 (average face)을 이용한 템플릿 매칭을 사용하였고, 실험을 통해서 제안한 방법이 효과적으로 수행됨을 보였다.

• 전기학회논문지
• /
• 제64권7호
• /
• pp.1104-1113
• /
• 2015

In this paper, we propose Two-Dimensional Robust Face Recognition System Realized with the Aid of Facial Symmetry with Illumination Variation. Preprocessing process is carried out to obtain mirror image which means new image rearranged by using difference between light and shade of right and left face based on a vertical axis of original face image. After image preprocessing, high dimensional image data is transformed to low-dimensional feature data through 2-directional and 2-dimensional Principal Component Analysis (2D)2PCA, which is one of dimensional reduction techniques. Polynomial-based Radial Basis Function Neural Network pattern classifier is used for face recognition. While FCM clustering is applied in the hidden layer, connection weights are defined as a linear polynomial function. In addition, the coefficients of linear function are learned through Weighted Least Square Estimation(WLSE). The Structural as well as parametric factors of the proposed classifier are optimized by using Particle Swarm Optimization(PSO). In the experiment, Yale B data is employed in order to confirm the advantage of the proposed methodology designed in the diverse illumination variation

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2000년도 제15차 학술회의논문집
• /
• pp.180-180
• /
• 2000

The shape of cold strip for the stainless process has been become issue in quality recently, and hence POSCO (Pohang Iron & Steel Co., Ltd) developed an automatic control system for strip shape in the sendzimir mill. The strip shape is measured by an outward measuring roll and is controlled by As_U roll and first intermediate roll. As_U roll consists of 8 saddles, which are controlled vertically. The fist intermediate rolls, which are controlled horizontally, consist of two pairs of rolls up and down. A developed shape control system is applied to real plant by using fuzzy logic and neural network method to control actuators; As_U roll and first intermediate roll. This system composes mainly of three parts as a real-time system, input to output conditioner board, and man-machine interface. The actual shape is recognized by neural network and converted into symmetric shape. The fuzzy controller, based on the shape from neural network and sensor, controls positions of the As_U roll and first intermediate roll. This paper verifies the shape controller performance. The experiments are made on line for the sendzimir mill. The shape control performance shows very efficient for the target tracking, shape symmetry, and fluctuation of shape.

• 대한원격탐사학회지
• /
• 제24권5호
• /
• pp.473-481
• /
• 2008

This paper presents an approach to Back-propagation and Radial Basis Function neural network method with various training set for automatic vehicle detection from aerial images. The initial extraction of candidate object is based on Mean-shift algorithm with symmetric property of a vehicle structure. By fusing the density and the symmetry, the method can remove the ambiguous objects and reduce the cost of processing in the next stage. To extract features from the detected object, we describe the object as a log-polar shape histogram using edge strengths of object and represent the orientation and distance from its center. The spatial histogram is used for calculating the momentum of object and compensating the direction of object. BPNN and RBFNN are applied to verify the object as a vehicle using a variety of non-car training sets. The proposed algorithm shows the results which are according to the training data. By comparing the training sets, advantages and disadvantages of them have been discussed.

• 한국정보처리학회논문지
• /
• 제6권12호
• /
• pp.3559-3568
• /
• 1999

수많은 정보의 홍수 속에 정보 과다는 현대인의 피할 수 없는 문제로 대두되었다. 특히, 인터넷과 컴퓨팅 기술의 발전으로 정보 자원이 급속도로 증가하고 있다. 따라서, 사용자가 원하는 정보를 찾아내는 것은 더욱 어려워졌다. 이러한 정보 검색 문제들을 해결하기 위하여 많은 정보 검색 시스템이 나타나게 되었다. 현재의 정보 검색 시스템들은 문서 검색에서는 사용자의 요구에 맞는 결과를 찾아 주고 있다. 그러나. 화상에 대한 검색 시스템의 연구는 초보 단계이기 때문에 사용자의 요구에 맞는 결과를 출력하지 못하고 있다. 이러한 문제를 해결하기 위해서 본 논문에서는 화상 검색 시스템이 사용자 모델을 이용하여 사용자에게 적응할 수 있는 기능을 부여하기 위하여 지능 사용자 인터페이스에 관한 고찰을 통한 인간-컴퓨터의 상호 작용 모델인 HCOS(Human-Computer Symmetry) 모델을 적용하고 이를 기반으로 화상 검색 시스템에서의 사용자 적응 능력을 갖는 지능 사용자 인터페이스를 제안하였다. 지능 사용자 인터페이스는 정보 검색의 복잡성과 사용자와 시스템간의 의미상의 차이를 감소시켜야 하며 사용자의 질의 성향과 관심을 반영할 수 있도록 학습 기능과 적응 기능을 포함하고 있어야 한다. 이를 위해 본 논문에서는 화상 검색을 위한 사용자 모델에 기계 학습(Machine Learning) 알고리듬인 결정 트리(Decision Tree)와 역전파 신경망(Backpropagation Neural Network)을 사용하였다. 지능 사용자 인터페이스의 화상 검색 실험을 통하여 시스템이 사용자에 적응하여 검색 효율이 좋아짐을 알 수 있었다.

• International Journal of Computer Science & Network Security
• /
• 제22권10호
• /
• pp.271-281
• /
• 2022

Manual seed classification challenges can be overcome using a reliable and autonomous seed purity identification and classification technique. It is a highly practical and commercially important requirement of the agricultural industry. Researchers can create a new data mining method with improved accuracy using current machine learning and artificial intelligence approaches. Seed classification can help with quality making, seed quality controller, and impurity identification. Seeds have traditionally been classified based on characteristics such as colour, shape, and texture. Generally, this is done by experts by visually examining each model, which is a very time-consuming and tedious task. This approach is simple to automate, making seed sorting far more efficient than manually inspecting them. Computer vision technologies based on machine learning (ML), symmetry, and, more specifically, convolutional neural networks (CNNs) have been widely used in related fields, resulting in greater labour efficiency in many cases. To sort a sample of 3000 seeds, KNN, SVM, CNN and CNN-SVM hybrid classification algorithms were used. A model that uses advanced deep learning techniques to categorise some well-known seeds is included in the proposed hybrid system. In most cases, the CNN-SVM model outperformed the comparable SVM and CNN models, demonstrating the effectiveness of utilising CNN-SVM to evaluate data. The findings of this research revealed that CNN-SVM could be used to analyse data with promising results. Future study should look into more seed kinds to expand the use of CNN-SVMs in data processing.