Extremely toughened Ti-based solid-solution carbide cermets

Kwak, Nojun,Kang, Shinhoo,Park, Choongkwon,Han, Heung Nam Elsevier 2019 JOURNAL OF ALLOYS AND COMPOUNDS Vol.784 No.-

<P><B>Abstract</B></P> <P>A new cermet with extremely high toughness of ∼22 MPa m<SUP>1/2</SUP> is designed. The high toughness of the cermet is mainly attributed to its characteristic microstructure, i.e., the even distribution of highly tough ceramic particles, unusual core/rim structure with a low internal strain, and strong bridging ligament. The microstructure and toughness of (Ti,W)C-Ni solid-solution cermets are strongly influenced by the addition of WC, i.e., by the W concentrations in the rim and binder phases. As the WC content increases, the shape of the carbide particles changes from faceted to rounded, and the toughness of the cermets increases sharply. In addition, crack propagation induced by an indentation is altered from the trans-granular mode to the inter-granular mode when the WC content is increased.</P> <P><B>Highlights</B></P> <P> <UL> <LI> We designed new cermet with exceptional toughness by microstructural control. </LI> <LI> Highest value was recorded as K<SUB>1C</SUB>:>21 MPa m<SUP>1/2</SUP> with a corresponding H<SUB>V</SUB>:>11 GPa. </LI> <LI> The properties were obtained by generating an unusual core/rim microstructure. </LI> <LI> The exploiting of (Ti<SUB>0.88</SUB>W<SUB>0.12</SUB>)C nanoparticles ≤30 nm was a key engineering idea. </LI> </UL> </P>

자율주행자동차를 위한 영상인식기술 동향

곽노준(Nojun Kwak),정지수(Jisoo Jeong) 한국자동차공학회 2018 오토저널 Vol.40 No.6

가상환경내에서 사용되는 Go - Go 기법들의 평가

곽노준(Nojun Kwak),강상우(Sangwoo Kang),송창근(ChangGuen Song) 한국정보과학회 1999 한국정보과학회 학술발표논문집 Vol.26 No.1B

가상환경은 사람과 컴퓨터간의 새로운 접근방식을 제공해 준다. 사용자는 더 이상 관찰자가 아니라 가상공간상의 적극적 참여자가 되는 것이다. 몰입형 가상환경 응용분야에서 상호작용기술의 중요성은 점점 더 증대되고 있다. 그러나 현재까지 개발된 상호작용기술은 효과적이지 못하고 사용하기 편리하지 못하기 때문에 가상현실 기술이 상용화되지 못하고 있다. 기존에 제안된 사용자중심 Metaphor기반의 Go-Go기법과 변형된 Go-Go기법들을 구현하고 20명의 피 실험자에 대한 실험을 통하여 각 기법의 특징 및 성능을 평가하고 얻어진 데이터를 분석하였다. 실험을 통해 일정한 Metaphor를 유지하는 기법은 어느 특정 작업(Task)에서만 효과적임을 알 수 있었다. 기존의 Go-Go기법의 단점인 회전동작을 보완하기 회전버튼을 이용한 기법을 구현하여 그 성능을 높였다. 효과적인 기법의 개발을 위해서는 서로 다른 기법과 상이한 Metaphor의 장점을 혼합하여 새로운 기법을 개발하고 신뢰성확보를 위한 정량적인 실험과 사용자대상의 사용성평가가 필요하다.

3D 애니메이션 제작을 위한 자동차 모션 분석 및 셋업 사례

곽노준(Kwak Nojun) 한국디지털디자인학회 2009 디지털디자인학연구 Vol.9 No.3

초기의 미국 애니메이션 모습을 살펴보면 캐릭터와 자동차의 형태를 단순화 시킨 것을 알 수 있다. 하지만 당시에 자동차는 사람이나 동물 캐릭터와 비교하여 상대적으로 중요하지 않은 캐릭터로 여겨졌으며 관련 기술 역시 크게 발전하지 못하였다. 최근에 급속하게 발전한 3D 애니메이션 기술은 복잡한 형태의 메카닉 캐릭터의 모션구현을 비교적 쉽게 제작할 수 있도록 하였으나 이론적 노하우의 정리 및 관련 데이터의 체계화가 이루어지지 않아 제작의 효율성에 있어 문제점이 제기되고 있다. 본 연구에서는 효과적인 모션 구현을 위하여 자동차의 구조적 특징을 살펴보고 3D 자동차 셋업 사례들을 소개 하고자 한다. 차체의 기본운동은 이동변위(translate)와 회전변위(rotate)로 나타낼 수 있으며 특히 회전변위의 피봇점은 차량의 무게 중심과 일치해야 한다. 휠의 회전운동은 간단한 수식을 이용하여 이동변위만으로 회전변위작업을 동시에 할 수 있다. 또한 현가장치의 경우도 수식과 스크립트를 이용하여 이동변위와 적절한 보정작업을 통하여 모션 작업을 구현할 수 있다. 자동차는 강체적 성격(rigid body)을 가지고 있기 때문에 그 모션 구현에 비교적 간단한 수식만을 이용하면서도 큰 효과를 볼 수 있다. 수식 이용의 효과는 몇 가지 셋업 사례들의 작업공정을 비교해 보았을 때 그 동작의 정확성과 효율성이 현격하게 증가함을 봄으로써 알 수 있다. In the early days of American animation animators should simplify the shape of character and car to raise the efficiency of production. At that time car was not considered as important character and related expression skill was not developed. Lately problem in efficiency of manufacture is advanced because 3D animation technology that develop festinately did to manufacture easily motion of complex Mechanic Character but retrenchment of theoretical know-how and systematization of connection data do not consist. The purpose of this paper is to propose optimized setup for 3D car and that will make a high degree of efficiency in 3D key animation process. Basis motion of car body can appear by transfer displacement (translate) and turning displacement (rotate). Pivot of turning displacement must agree with center of gravity of vehicles specially. Turn of wheel can do turning displacement work at the same time by transfer displacement using simple formula. Also in case of suspension can make motion work through transfer displacement and suitable revision work because using numerical formula and script. A car can see big effect using comparative simple expression in the motion embodiment because have a personality of rigid body. Effect of numerical formula utilization can know seeing that accuracy and efficiency of these action increase as is different when compared work processes of some Setup examples.

Implementing Kernel Methods Incrementally by Incremental Nonlinear Projection Trick

IEEE 2017 IEEE transactions on cybernetics Vol.47 No.11

<P>Recently, the nonlinear projection trick (NPT) was introduced enabling direct computation of coordinates of samples in a reproducing kernel Hilbert space. With NPT, any machine learning algorithm can be extended to a kernel version without relying on the so called kernel trick. However, NPT is inherently difficult to be implemented incrementally because an ever increasing kernel matrix should be treated as additional training samples are introduced. In this paper, an incremental version of the NPT (INPT) is proposed based on the observation that the centerization step in NPT is unnecessary. Because the proposed INPT does not change the coordinates of the old data, the coordinates obtained by INPT can directly be used in any incremental methods to implement a kernel version of the incremental methods. The effectiveness of the INPT is shown by applying it to implement incremental versions of kernel methods such as, kernel singular value decomposition, kernel principal component analysis, and kernel discriminant analysis which are utilized for problems of kernel matrix reconstruction, letter classification, and face image retrieval, respectively.</P>

Nonlinear Projection Trick in Kernel Methods: An Alternative to the Kernel Trick

IEEE 2013 IEEE transactions on neural networks and learning Vol.24 No.12

<P>In kernel methods such as kernel principal component analysis (PCA) and support vector machines, the so called kernel trick is used to avoid direct calculations in a high (virtually infinite) dimensional kernel space. In this brief, based on the fact that the effective dimensionality of a kernel space is less than the number of training samples, we propose an alternative to the kernel trick that explicitly maps the input data into a reduced dimensional kernel space. This is easily obtained by the eigenvalue decomposition of the kernel matrix. The proposed method is named as the nonlinear projection trick in contrast to the kernel trick. With this technique, the applicability of the kernel methods is widened to arbitrary algorithms that do not use the dot product. The equivalence between the kernel trick and the nonlinear projection trick is shown for several conventional kernel methods. In addition, we extend PCA-L1, which uses L<SUB>1</SUB>-norm instead of L<SUB>2</SUB>-norm (or dot product), into a kernel version and show the effectiveness of the proposed approach.</P>

Time-Domain Measurement Data Accumulation for Slow Moving Point Target Detection in Heavily Cluttered Environments Using CNN

조원민,Nojun Kwak 한국전자파학회 2023 Journal of Electromagnetic Engineering and Science Vol.23 No.6

In modern radars, the target detection probability is increased by lowering the detection threshold via signal processing to detect a point target with a small radar cross-section value. However, a lower threshold increases the number of false targets. In the conventional tracking method, which uses a general tracking filter, the measurement data between scans should be compared. Therefore, for a large amount of acquired measurement data, the computational complexity can be reduced by accumulating the acquired measurement data over time, recognizing the target movement as a pattern, and training a convolutional neural network (CNN) model. Here, we propose a method to create a desired target scenario by transfer learning and estimate the target position using the activation map of a binary detector CNN model. The model can detect a target using the actual acquired radar data, and the processing time remains constant, regardless of the number of false alarms.

Independent component analysis by lp-norm optimization

Park, Sungheon,Kwak, Nojun Elsevier 2018 Pattern recognition Vol.76 No.-

<P><B>Abstract</B></P> <P>In this paper, a couple of new algorithms for independent component analysis (ICA) are proposed. In the proposed methods, the independent sources are assumed to follow a predefined distribution of the form f ( s ) = α exp ( − β | s <SUP> | p </SUP> ) and a maximum likelihood estimation is used to separate the sources. In the first method, a gradient ascent method is used for the maximum likelihood estimation, while in the second, a non-iterative algorithm is proposed based on the relaxation of the problem. The maximization of the log-likelihood of the estimated source <I>X<SUP>T</SUP> </I> <B>w</B> given the parameter <I>p</I> and the data <I>X</I> is shown to be equivalent to the minimization of <I>l<SUB>p</SUB> </I>-norm of the projected data <I>X<SUP>T</SUP> </I> <B>w</B>. This formulation of ICA has a very close relationship with the Lp-PCA where the maximization of the same objective function is solved. The proposed algorithm solves an approximation of the <I>l<SUB>p</SUB> </I>-norm minimization problem for both super-(<I>p</I> < 2) and sub-Gaussian (<I>p</I> > 2) cases and shows superior performance in separating independent sources than the state of the art algorithms for ICA computation.</P> <P><B>Highlights</B></P> <P> <UL> <LI> We propose a couple of algorithms for independent component analysis using lp-norm optimization. </LI> <LI> By the use of the generalized Gaussian function, the problem of ICA was shown to be equivalent to the optimization of lp-norm. </LI> <LI> The proposed method is robust to the outliers especially for super-Gaussian sources. </LI> </UL> </P>

Ductility enhancement of tungsten after plastic deformation

Oh, Yeonju,Kwak, Nojun,Lee, Keunho,Ko, Won-Seok,Han, Heung Nam ELSEVIER SCIENCE 2019 Journal of Alloys and Compounds Vol.787 No.-

<P><B>Abstract</B></P> <P>An unusual room temperature mechanical behavior of pure tungsten is investigated by focusing on three specimens prepared with different microstructures: as-received (hot-rolled), recrystallized, and cold-rolled specimens. Contrary to ordinary expectations in metallic materials, only the cold-rolled specimen exhibits significant plastic deformation during tensile testing, with improved strength and ductility, while the recrystallized and the as-received specimens fail in the elastic region. We further provide an explanation of such characteristics by systematically utilizing experimental and theoretical analysis at a small scale: nano-indentation tests clarify the inherent mechanical response inside grains and provide a statistical distribution of the maximum shear stress during pop-in events, corresponding to onset of plastic yielding; atomistic simulations provide information on the overall fracture strength of various grain boundaries. By comparing the maximum shear stress and the grain boundary fracture stress, we could explain that the distinctive plasticity in the cold-rolled specimen is caused by the movement of pre-existing dislocations before grain boundary fracture. This contrasts the recrystallized and as-received specimens, where grain boundary fracture occurs prior to the nucleation of dislocations or activation of dislocation sources.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Distinctive plasticity of cold rolled tungsten was investigated at room temperature. </LI> <LI> The more pre-existing dislocations, the lower shear stress for plastic yielding. </LI> <LI> Molecular dynamic simulations provide the fracture stress of grain boundaries, ∼16 GPa. </LI> <LI> Cold rolled tungsten has lower stress for plastic yielding than the fracture stress. </LI> <LI> Ductility was enhanced by the dislocation activation prior to intergranular fracture. </LI> </UL> </P> <P><B>Graphical abstract</B></P> <P>[DISPLAY OMISSION]</P>

회귀문제를 위한 비선형 특징 추출 방법

김성민(Kim Seongmin),곽노준(Kwak Nojun) 한국방송·미디어공학회 2010 한국방송공학회 학술발표대회 논문집 Vol.2010 No.11

본 논문에서는 회귀문제를 위한 비선형 특징 추출방법을 제안하고 분류문제에 적용한다. 이 방법은 이미 제안된 선형판별 분석법을 회귀문제에 적용한 회귀선형판별분석법(Linear Discriminant Analysis for regression:LDAr)을 비선형 문제에 대해 확장한 것이다. 본 논문에서는 이를 위해 커널함수를 이용하여 비선형 문제로 확장하였다. 기본적인 아이디어는 입력 특징 공간을 커널 함수를 이용하여 새로운 고차원의 특징 공간으로 확장을 한 후, 샘플 간의 거리가 큰 것과 작은 것의 비율을 최대화하는 것이다. 일반적으로 얼굴 인식과 같은 응용 분야에서 얼굴의 크기, 회전과 같은 것들은 회귀문제에 있어서 비선형적이며 복잡한 문제로 인식되고 있다. 본 논문에서는 회귀 문제에 대한 간단한 실험을 수행하였으며 회귀선형판별분석법(LDAr)을 이용한 결과보다 향상된 결과를 얻을 수 있었다.