• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.38 no.9
• /
• pp.933-941
• /
• 2014

This paper provides a comparison between the Krylov subspace method (KSM) and modal truncation method (MTM), which are typical projection-based model order reduction methods. The frequency responses are compared to determine the numerical accuracies and efficiencies. In order to compare the numerical accuracies of the KSM and MTM, the frequency responses and relative errors according to the order of the reduced model and frequency of interest are studied. Subsequently, a numerical examination shows whether a reduced order can be determined automatically with the help of an error convergence indicator. As for the numerical efficiency, the computation time needed to generate the projection matrix and the solution time to perform a frequency response analysis are compared according to the reduced order. A finite element model for a car suspension is considered as an application example of the numerical comparison.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.7
• /
• pp.2599-2613
• /
• 2015

Multi-index hashing (MIH) is the state-of-the-art method for indexing binary codes, as it di-vides long codes into substrings and builds multiple hash tables. However, MIH is based on the dataset codes uniform distribution assumption, and will lose efficiency in dealing with non-uniformly distributed codes. Besides, there are lots of results sharing the same Hamming distance to a query, which makes the distance measure ambiguous. In this paper, we propose a data-oriented multi-index hashing method (DOMIH). We first compute the covariance ma-trix of bits and learn adaptive projection vector for each binary substring. Instead of using substrings as direct indices into hash tables, we project them with corresponding projection vectors to generate new indices. With adaptive projection, the indices in each hash table are near uniformly distributed. Then with covariance matrix, we propose a ranking method for the binary codes. By assigning different bit-level weights to different bits, the returned bina-ry codes are ranked at a finer-grained binary code level. Experiments conducted on reference large scale datasets show that compared to MIH the time performance of DOMIH can be improved by 36.9%-87.4%, and the search accuracy can be improved by 22.2%. To pinpoint the potential of DOMIH, we further use near-duplicate image retrieval as examples to show the applications and the good performance of our method.

• Journal of KIISE:Software and Applications
• /
• v.33 no.5
• /
• pp.481-488
• /
• 2006

The performance of face recognition methods using subspace projection is directly related to the characteristics of their basis images, especially in the cases of local distortion or partial occlusion. In order for a subspace projection method to be robust to local distortion and partial occlusion, the basis images generated by the method should exhibit a part-based local representation. We propose an effective part-based local representation method named locally salient ICA (LS-ICA) method for face recognition that is robust to local distortion and partial occlusion. The LS-ICA method only employs locally salient information from important facial parts in order to maximize the benefit of applying the idea of 'recognition by parts.' It creates part-based local basis images by imposing additional localization constraint in the process of computing ICA architecture I basis images. We have contrasted the LS-ICA method with other part-based representations such as LNMF (Localized Non-negative Matrix Factorization) and LFA (Local Feature Analysis). Experimental results show that the LS-ICA method performs better than PCA, ICA architecture I, ICA architectureII, LFA, and LNMF methods, especially in the cases of partial occlusions and local distortions.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.11
• /
• pp.1080-1087
• /
• 2009

This paper presents an algorithm to acquire 3D geometric information using a virtual plane method. The method to measure 3D information on the plane is easy, because it's not concerning value on the z-axis. A plane can be made by arbitrary three points in the 3D space, so the algorithm is able to make a number of virtual planes from feature points on the target object. In this case, these geometric relations between the origin of each virtual plane and the origin of the target object coordinates should be expressed as known homogeneous matrices. To include this idea, the algorithm could induce simple matrix formula which is only concerning unknown geometric relation between the origin of target object and the origin of camera coordinates. Therefore, it's more fast and simple than other methods. For achieving the proposed method, a regular pin-hole camera model and a perspective projection matrix which is defined by a geometric relation between each coordinate system is used. In the final part of this paper, we demonstrate the techniques for a variety of applications, including measurements in industrial parts and known patches images.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.62 no.8
• /
• pp.1151-1156
• /
• 2013

In this study, to clearly establish the concept of a geometric modeling I apply for the concept of Pushbroom, limited to two-dimensional radiation Locator to provide a three-dimensional information purposes. Respect to the radiation scanner Pushbroom modeling techniques, geometric modeling method was presented introduced to extract three-dimensional information as long as the rotational component of the Gamma-Ray Linear Pushbroom Stereo System, introduced the two-dimensional and three-dimensional spatial information in the matching relation that can be induced. In addition, the pseudo-inverse matrix by using the conventional least-squares method, GCP(Ground Control Point) to demonstrate compliance by calculating the key parameters. Projection transformation matrix is calculated for obtaining three-dimensional information from two-dimensional information can be used as the primary relationship, and through the application of a radiation image matching technology will make it possible to extract three-dimensional information from two-dimensional X-ray imaging.

• Journal of KIISE:Software and Applications
• /
• v.32 no.8
• /
• pp.789-794
• /
• 2005

LDA (Linear Discriminant Analysis) provides the projection that discriminates the data well, and shows a very good performance for face recognition. However, since LDA provides only one transformation matrix over whole data, it is not sufficient to discriminate the complex data consisting of many classes like honan faces. To overcome this weakness, we propose a new face recognition method, called LDA mixture model, that the set of alf classes are partitioned into several clusters and we get a transformation matrix for each cluster. This detailed representation will improve the classification performance greatly. In the simulation of face recognition, LDA mixture model outperforms PCA, LDA, and PCA mixture model in terms of classification performance.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.1
• /
• pp.31-39
• /
• 2015

This paper deals with a method for estimating variance components on the basis of projections under the assumption of random effects model. It discusses how to use projections for getting sums of squares to estimate variance components. The use of projections makes the vector subspace generated by the model matrix to be decomposed into subspaces that are orthogonal each other. To partition the vector space by the model matrix stepwise procedure is used. It is shown that the suggested method is useful for obtaining Type I sum of squares requisite for the ANOVA method.

• Journal of Korea Foundry Society
• /
• v.13 no.3
• /
• pp.238-247
• /
• 1993

Microstructure, hardness and wear characteristics of $SiC_p/Al-6.5wt%Si-1.7wt%Mg$ alloy composites fabricated by the method of rheo-compocasting and hot pressing are investigated in this study. The dispersion of SiC particles in the composites is homogeneous and the hardness improves as additional amount increases. The wear amount of the matrix metal increases highly as wear rates increase, for the wear mechanism changes from adhesive wear to melt wear, and the matrix metal was coated on the surface of revolving disc and its weight increases. In the 5vol% composites, Fe is adhered on the surface of specimen by the projection of the dispersed hard SiC particles which have net-work structure and the coating layer is about $300{\mu}m$. But in the composite more than 20vol%, the wear amount of composite decreases because the SiC particles which have superior hardness, wear resistance and heat resistance properties resist wear, the abrasive wear turn out predominant wear mechanism and so the wear amount of revolving disc increases.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.3
• /
• pp.1243-1263
• /
• 2018

The two-stage linear discrimination analysis (TSLDA) is a feature extraction technique to solve the small size sample problem in the field of image recognition. The TSLDA has retained all subspace information of the between-class scatter and within-class scatter. However, the feature information in the four subspaces may not be entirely beneficial for classification, and the regularization procedure for eliminating singular metrics in TSLDA has higher time complexity. In order to address these drawbacks, this paper proposes an improved two-stage linear discriminant analysis (Improved TSLDA). The Improved TSLDA proposes a selection and compression method to extract superior feature information from the four subspaces to constitute optimal projection space, where it defines a single Fisher criterion to measure the importance of single feature vector. Meanwhile, Improved TSLDA also applies an approximation matrix method to eliminate the singular matrices and reduce its time complexity. This paper presents comparative experiments on five face databases and one handwritten digit database to validate the effectiveness of the Improved TSLDA.

• Spatial Information Research
• /
• v.5 no.2
• /
• pp.177-183
• /
• 1997

S1he concept of a class in the land-use classification system can be equally applied to a class in the land-use-change classification. The maximum likelihood method using linear discriminant function and Markov transition matrix method were integrated to a synthetic modeling effort in order to project spatial allocation of land-use-change and quantitative assignment of that prediction as a whole. The algorithm of both the multivariate discriminant function and the Markov chain matrix were discussed and the test of synthetic model on the study area was resulted in the projection of '90 year as well as '95 year land -use classification. The accuracy and the issue of modeling improvement were discussed eventually.