• Journal of Information Processing Systems
• /
• v.14 no.1
• /
• pp.150-161
• /
• 2018

We propose a novel deep neural network model for detecting human activities in untrimmed videos. The process of human activity detection in a video involves two steps: a step to extract features that are effective in recognizing human activities in a long untrimmed video, followed by a step to detect human activities from those extracted features. To extract the rich features from video segments that could express unique patterns for each activity, we employ two different convolutional neural network models, C3D and I-ResNet. For detecting human activities from the sequence of extracted feature vectors, we use BLSTM, a bi-directional recurrent neural network model. By conducting experiments with ActivityNet 200, a large-scale benchmark dataset, we show the high performance of the proposed DeepAct model.

• Journal of KSNVE
• /
• v.9 no.2
• /
• pp.377-386
• /
• 1999

This paper is a study on model method for estimating system inputs from vibration responses, which is one of indirect input identification methods in frequency domain. The method has advantages over direct inverse method especially when points of operational inputs are inaccessible so that artificial excitation forces cannot be applied to obtain frequency response functions of the complete system. Procedures of extended modal model method are proposed and checked by numerical experiment. Mechanisms of error propagation, i.e., how errors in modal parameters such as poles nad mode shape vectors affect estimation of the input forces, are illustrated. Then, in order to counteract the error propagation, discrete modal filter approach is taken in this paper to compute the inversion of modal matrix in which the most serious errors seem to be generated. Further, a Reduced form of Modified Reciprocal Modal Vector(RMRMV) is proposed for estimating multiple inputs. It is shown to have smaller orthogonality error than MRMV.

• Journal of Power Electronics
• /
• v.17 no.6
• /
• pp.1480-1489
• /
• 2017

In this study, a carrier-based pulse-width modulation (PWM) method for two-level voltage source inverters in the over-modulation region is proposed. Based on the superposition principle, the reference voltage vectors outside the linear modulation boundary are adjusted to relocate to the vector hexagon, while their fundamental magnitudes are retained. In accordance with the adjusted reference vector, the corresponding modulated waves are respectively deduced in over-modulation mode I and II to generate the gate signals of the power switches, guaranteeing the linearity of the fundamental output phase voltage in the over-modulation region. Moreover, due to the linear relationship between the voltage vector and the duty ratios, the complicated sector identification and holding angle calculation found in previous methods are avoided in the modulated wave synthesis, which provides great simplicity for the proposed carrier-based over-modulation strategy. Experimental results demonstrate the effectiveness and validity of the proposed method.

• Journal of Species Research
• /
• v.11 no.3
• /
• pp.208-237
• /
• 2022

A total of 69 genera and 202 species of the family Syrphidae have been previously recognized in Korea. In the present study, we report the following 10 species new to Korea: Epistrophe aeka Kimura, 1989, Epistrophe aino (Matsumura, 1917), Epistrophe annulitarsis(Stackelberg, 1918), Epistrophe griseofasciata (Matsumura, 1918), Epistrophe sasayamana (Matsumura, 1918), Epistrophe shibakawae (Matsumura, 1917), Fagisyrphus cinctus (Fallén, 1817), Melangyna olsufjevi (Violovitsh, 1956), Melangyna pavlovskyi (Violovitsh, 1956), and Paragus clausseni Mutin, 1999. Among the four genera, the genus Fagisyrphus Dušek and Láska, 1967 is recognized for the first time in Korea. As a result, 70 genera and 212 species of the family Syrphidae are now officially recognized for the Korean fauna. For each newly recognized Korean species, we provide synonymy, diagnosis, and description with color photographs. In addition, we provide new Korean names for all newly recorded species as well as the genera Epistrophe, Fagisyrphus, and Melangyna, which previously did not have Korean names.

• Journal of the Korean Society for Precision Engineering
• /
• v.10 no.4
• /
• pp.54-63
• /
• 1993

This paper presents the dynamic modeling and control methodology to arrest structural deflections of industrial robotic manipulators featuring elastic members retrofitted with surface bonded pizoelectric actuators and sensors. The cynamic modeling is accomplished by employing a variational theorem, prior to developing a finite element formulation. This finite element formulation accounts for both original robot member elements and also bonded piezoelectric material elements. The governing equation of motion is then modified by condensing the electric potential vectors and subsequently two different negative velocity feedback controllers are established; a constant-gain feedback controller and a constant- amplitude feedback controller. By adopting a Model P50 articulating industrial robot manufactured by Gerneral Electric Company, conputer simulations are underlaken in order to demonstrate superior performance characteristics to be accrued from this proposed methodology such as smaller deflections at the end-effector.

• Bulletin of the Korean Mathematical Society
• /
• v.24 no.1
• /
• pp.13-17
• /
• 1987

Let S denote the class of nivalent functions normalized so that f(0)=f'(0)-1=0 in vertical bar z vertical bar <1. Let $S_{\alpha}$$^{*}$, -.pi./2<.alpha.<.pi./2, denote the subclass of S that satisfies Re $e^{i{\alpha}}$zf'(z)/f(z).geq.0 in vertical bar z vertical bar <1; then f is called .alpha.-spiral-like and the case .alpha.=0 is the class of normalized starlike functions [6, pp.52]. Let T denote the class of functions f normalized as above and satisfying Im z[Im f(z)]..geq.0 in vertical bar z vertical bar <1; then f is called typically real and T contains those functions of S whose coefficients are real [6, pp.55]. Also, in view of [6, pp.231], let B(.lambda.) be the class of function normalized as above and map vertical bar z vertical bar <1 onto the complement of an arc with radial angle .lambda.(0<.lambda.<.pi./2). The radial angle is meant to be the angle between the tangent and radial vectors to the arc. This note includes a sharp version for Corollary 1 of [2] when f.mem. $S_{\alpha}$$^{*}$ as well as a logarithmic coefficient estimate.nt estimate.

• Journal of English Language & Literature
• /
• v.56 no.6
• /
• pp.1135-1161
• /
• 2010

The interpretations of degree expressions require the postulation of new entities to represent degrees. Diverse entities such as degrees, intervals, and vectors are adopted for degree expressions. Positive degree sentences and questions are properly construed with the introduction of these entities, but their negative counterparts need more consideration. Negative degree sentences show dual patterns of entailments depending on contexts, and negative degree questions are unacceptable, making weak islands. To explicate the distinct nature of negative degree sentences and questions, Fox & Hackl (2006) provide an analysis based on degrees while Abrusan & Spector (2010) suggest a proposal in interval readings of degree expressions. I have pointed out the theoretical problems of these analyses and proposed an alternative in the framework of the vector space semantics, following Winter (2005). Bi-directional scales in vector space fit well with the dual patterns of negative degree sentences, and the notion of a reference vector is useful to accommodate the contextual influence in negative degree sentences and to deal with the unacceptability of negative degree questions.

• Korea Trade Review
• /
• v.45 no.4
• /
• pp.95-105
• /
• 2020

Vietnam has experienced a high economic growth since early 2000s. One of the reasons for this successful economic growth is foreign direct investment that has been invested mainly in manufacturer sector in Vietnam. In this paper, we examine the impacts of foreign direct investment to Vietnam on its exports using quarterly data from 2000:1 to 2017:4. Since all the variables in our model is subject to I(1), we apply Fully Modified OLS(FMOLS) to estimate a cointegration vectors. Our results show that there exists a long-run relationship among Export, FDI, Exchange rate and G20 countries' GDP. Also, we find that FDI has a positive effect on Vietnam's export, which was statistically significant. Our results support the hypothesis that the FDI to Vietnam since 2000 has an export-oriented feature.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.7
• /
• pp.4936-4941
• /
• 2015

This study(I) has analyzed hydraulic characteristics with pier shapes by the bridge construction. The pier shapes are classified into total six types such as square, rhombus, octagon, oval, round, and no-piers. One-dimensional model(HEC-RAS) and two-dimensional model (RMA-2) were employed to analyze hydraulic characteristics around bridge piers. Square and rhombus shapes of piers showed velocity vectors in the upstream direction, which has a significant impact on the river bed changes by erosion and sediment transport around the piers. The flow characteristics of the oval type pier was most similar to that of no-pier situation almost without disrupting the river flow. This analysis can help to select pier types in the new bridge construction for the future.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.27 no.5
• /
• pp.383-392
• /
• 2007

In this paper, we study the prediction of depth and width of a defect in steam generator tube in nuclear power plant using neural network. To this end, we first generate eddy current testing (ECT) signals for 4 defect patterns of SG tube: I-In type, I-Out type, V-In type, and V-Out type. In particular, we generate 400 ECT signals for various widths and depths for each defect type by the numerical analysis program based on finite element modeling. From those generated ECT signals, we extract new feature vectors for the prediction of defect size, which include the angle between the two points where the maximum impedance and half the maximum impedance are achieved. Using the extracted feature vector, multi-layer perceptron with one hidden layer is used to predict the size of defects. Through the computer simulation study, it is shown that the proposed method achieves decent prediction performance in terms of maximum error and mean absolute percentage error (MAPE).