• Smart Media Journal
• /
• v.10 no.1
• /
• pp.16-24
• /
• 2021

The exiting word embedding methodology such as word2vec represents words, which only occur in the raw training corpus, as a fixed-length vector into a continuous vector space, so when mapping the words incorporated in the raw training corpus into a fixed-length vector in morphologically rich language, out-of-vocabulary (OOV) problem often happens. Even for sentence embedding, when representing the meaning of a sentence as a fixed-length vector by synthesizing word vectors constituting a sentence, OOV words make it challenging to meaningfully represent a sentence into a fixed-length vector. In particular, since the agglutinative language, the Korean has a morphological characteristic to integrate lexical morpheme and grammatical morpheme, handling OOV words is an important factor in improving performance. In this paper, we propose parallel Tri-LSTM sentence embedding that is robust to the OOV problem by extending utilizing the morphological information of words into sentence-level. As a result of the sentiment analysis task with corpus in Korean, we empirically found that the character unit is better than the morpheme unit as an embedding unit for Korean sentence embedding. We achieved 86.17% accuracy on the sentiment analysis task with the parallel bidirectional Tri-LSTM sentence encoder.

• Reproductive and Developmental Biology
• /
• v.32 no.3
• /
• pp.205-209
• /
• 2008

This study was carried out to develop knock-in vector for EGFP (enhanced green fluorescent protein) expression in porcine $\beta$-casein locus. For construction of knock-in vector using porcine $\beta$-casein gene, we cloned the $\beta$-casein genome DNA from porcine fetal fibroblast cells, EGFP and SV40 polyA signal using PCR. The knock-in vectors consisted of a 5-kb fragment as the 5' recombination arm and a 2.7-kb fragment as the 3' recombination arm. We used the neomycin resistance gene ($neo^{r}$) as a positive selectable marker and the diphtheria toxin A (DT-A) gene as a negative selectable marker. To demonstrate EGFP expression from knock-in vector, we are transfected knock-in vector that has EGFP gene in murine mammary epithelial cell line HC11 cells with pSV2 neo plasmid. The EGFP expression was detected in HC11 cells transfected knock-in vector. This result demonstrates that this knock-in vector may be used for the development of knock-in transgenic pig.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.5 s.305
• /
• pp.1-10
• /
• 2005

Content-based image retrieval systems are being actively investigated owing to their ability to retrieve images based on the actual visual content rather than by manually associated textual descriptions. In this paper, we propose a novel approach for image retrieval based on edge structural features using edge correlogram and color coherence vector. After color vector angle is applied in the pre-processing stage, an image is divided into two image parts (high frequency image and low frequency image). In low frequency image, the global color distribution of smooth pixels is extracted by color coherence vector, thereby incorporating spatial information into the proposed color descriptor. Meanwhile, in high frequency image, the distribution of the gray pairs at an edge is extracted by edge correlogram. Since the proposed algorithm includes the spatial and edge information between colors, it can robustly reduce the effect of the significant change in appearance and shape in image analysis. The proposed method provides a simple and flexible description for the image with complex scene in terms of structural features of the image contents. Experimental evidence suggests that our algorithm outperforms the recently histogram refinement methods for image indexing and retrieval. To index the multidimensional feature vectors, we use R*-tree structure.

• Journal of KSNVE
• /
• v.8 no.5
• /
• pp.949-956
• /
• 1998

Complex and large lattice type structures are frequently used in design of bridge, tower, crane and aerospace structures. In general, in order to analyze these structures we have used the finite element method(FEM). This method is the most widely used and powerful method for structural analysis lately. However, it is necessary to use a large amount of computer memory and computational time because the FEM requires many degrees of freedom for solving dynamic problems exactly for these complex and large structures. For analyzing these structures on a personal computer, the authors developed the transfer stiffness coefficient method(TSCM). This method is based on the concept of the transfer of the nodal dynamic stiffness coefficient matrix which is related to force and displacement vector at each node. And we suggested TSCM for free vibration analysis of complex and large lattice type structures in the previous report. In this paper, we formulate forced vibration analysis algorithm for complex and large lattice type structures using extened TSCM. And we confirmed the validity of TSCM through computational results by the FEM and TSCM, and experimental results for lattice type structures with harmonic excitation.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.56 no.12
• /
• pp.2135-2144
• /
• 2007

This paper presents system design of the slotless double-side Permanent Magnet Linear Synchronous Machine system (PMLSM) through magnetic field analysis and dynamic modeling. In our analysis, 2-D analytical treatments based on the magnetic vector potential were adopted to predict magnetic field with space harmonics by PM mover magnetization and stator winding current. From these, the design parameters such as inductance, Back-emf, and thrust are estimated. And, the electrical dynamic modeling including synchronous speed is completed by calculation of a DC link voltage in effort to obtain the accurate mechanical power from Space Vector Pulse Width Modulation(SVPWM). Therefore, the system design of PMLSM is performed from estimation of design parameters according to PM size and coil turns in magnetic field and from calculation of a DC link voltage to satisfy base speed and base thrust represented as the maximum output power in dynamic modeling. The estimated values from the analysis are verified by the finite element method and experimental results.

• Journal of the Korean Society of Safety
• /
• v.32 no.5
• /
• pp.179-185
• /
• 2017

Accident analysis is an essential process to make basic data for accident prevention. Most researches depend on survey data and accident statistics to analyze accidents, but these kinds of data are not sufficient for systematic and detailed analysis. We, in this paper, propose an accident classification model that extracts task type, original cause materials, accident type, and the number of deaths from accident reports. The classification model is a support vector machine (SVM) with word occurrence features, and these features are selected based on mutual information. Experiment shows that the proposed model can extract task type, original cause materials, accident type, and the number of deaths with almost 100% accuracy. We also develop an accident ontology to express the information extracted by the classification model. Finally, we illustrate how the proposed classification model and ontology effectively works for the accident analysis. The classification model and ontology are expected to effectively analyze various accidents.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.10
• /
• pp.2448-2463
• /
• 2013

The traditional block-based motion compensation methods in frame rate up-conversion (FRUC) only use a single uniquely motion vector field. However, there will always be some mistakes in the motion vector field whether the advanced motion estimation (ME) and motion vector analysis (MA) algorithms are performed or not. Once the motion vector field has many mistakes, the quality of the interpolated frame is severely affected. In order to solve the problem, this paper proposes a novel joint overlapped block motion compensation method (8J-OBMC) which adopts motion vectors of the interpolated block and its 8-neighbor blocks to jointly interpolate the target block. Since the smoothness of motion filed makes the motion vectors of 8-neighbor blocks around the interpolated block quite close to the true motion vector of the interpolated block, the proposed compensation algorithm has the better fault-tolerant capability than traditional ones. Besides, the annoying blocking artifacts can also be effectively suppressed by using overlapped blocks. Experimental results show that the proposed method is not only robust to motion vectors estimated wrongly, but also can to reduce blocking artifacts in comparison with existing popular compensation methods.

• Parasites, Hosts and Diseases
• /
• v.55 no.1
• /
• pp.55-60
• /
• 2017

Fascioliasis is a foodborne zoonotic parasitic disease. We report 4 cases occurring in the same family, in whom diagnosis of acute fascioliasis was established after series of tests. One case was hospitalized with fever, eosinophilia, and hepatic lesions. MRI showed hypodense changes in both liver lobes. The remaining 3 cases presented with the symptom of stomachache only. Stool analysis was positive for Fasciola eggs in 2 adult patients. The immunological test and molecular identification of eggs were confirmed at the National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention, Shanghai, China. The results of serological detection were positive in all the 4 patients. DNA sequencing of PCR products of the eggs demonstrated 100% homology with ITS and cox1 of Fasciola hepatica. The conditions of the patients were not improved by broad-spectrum anti-parasitic drugs until administration of triclabendazole.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.7
• /
• pp.37-45
• /
• 2013

Context dependent fusion (CDF) is a fusion algorithm that combines multiple outputs from different classifiers to achieve better performance. CDF tries to divide the problem context into several homogeneous sub-contexts and to fuse data locally with respect to each sub-context. CDF showed better performance than existing methods, however, it is sensitive to noise due to the large number of parameters optimized and the innate linearity limits the application of CDF. In this paper, a variant of CDF using support vector machines (SVMs) for fusion and kernel principal component analysis (K-PCA) for context extraction is proposed to solve the problems in CDF, named CDF-SVM. Kernel PCA can shape irregular clusters including elliptical ones through the non-linear kernel transformation and SVM can draw a non-linear decision boundary. Regularization terms is also included in the objective function of CDF-SVM to mitigate the noise sensitivity in CDF. CDF-SVM showed better performance than CDF and its variants, which is demonstrated through the experiments with a landmine data set.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.15 no.2
• /
• pp.209-218
• /
• 2002

The Vector Random Decrement technique has been developed as an efficient method for transforming ambient responses into free decays of linear structures. It is shown that the VRD functions nay contain as much information about the modes as the really measured free decay responses. In this paper, the theory of the VRD technique is extended by applying the concept of the mode shape ratio into the Ibrahim Time Domain modal parameter identification algorithm. The VRD function is not shifted in the correction procedures for constant time shifts of the proposed VRD technique. Thus, a number of points equal to the largest of the time shifts used in the vector triggering condition are not deleted. In the VRD functions, any influence of the input to the system is averaged out. The proposed technique is compared with the traditional VRD technique by assessment of the modal parameters. The applicability of the VRD technique has been justified through a simulation study and a study of the response of a laboratory beam model subject to ambient loads.