• 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.

• International Journal of Reliability and Applications
• /
• v.13 no.1
• /
• pp.19-35
• /
• 2012

This paper presents an optimum design of step-stress partially accelerated life test (PALT) plan which allows the test condition to be changed from use to accelerated condition on the occurrence of fixed number of failures. Various life distribution models such as exponential, Weibull, log-logistic, Burr type-Xii, etc have been used in the literature to analyze the PALT data. The need of different life distribution models is necessitated as in the presence of a limited source of data as typically occurs with modern devices having high reliability, the use of correct life distribution model helps in preventing the choice of unnecessary and expensive planned replacements. Truncated distributions arise when sample selection is not possible in some sub-region of sample space. In this paper it is assumed that the lifetimes of the items follow Truncated Logistic distribution truncated at point zero since time to failure of an item cannot be negative. Optimum step-stress PALT plan that finds the optimal proportion of units failed at normal use condition is determined by using the D-optimality criterion. The method developed has been explained using a numerical example. Sensitivity analysis and comparative study have also been carried out.

• Smart Structures and Systems
• /
• v.12 no.3_4
• /
• pp.235-250
• /
• 2013

Optimal sensor placement (OSP) technique plays a key role in the structural health monitoring (SHM) of large-scale structures. According to the mathematical background and implicit assumptions made in the triaxial effective independence (EfI) method, this paper presents a novel multi-dimensional OSP method for the Canton Tower focusing on application demands. In contrast to existing methods, the presented method renders the corresponding target mode shape partitions as linearly independent as possible and, at the same time, maintains the stability of the modal matrix in the iteration process. The modal assurance criterion (MAC), determinant of the Fisher Information Matrix (FIM) and condition number of the FIM have been taken as the optimal criteria, respectively, to demonstrate the feasibility and effectiveness of the proposed method. Numerical investigations suggest that the proposed method outperforms the original EfI method in all instances as expected, which is looked forward to be even more pronounced should it be used for other multi-dimensional optimization problems.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.11
• /
• pp.5496-5521
• /
• 2018

Object detection and tracking is the basic capability of mobile robots to achieve natural human-robot interaction. In this paper, an object tracking system of mobile robot is designed and validated using improved multiple instance learning algorithm. The improved multiple instance learning algorithm which prevents model drift significantly. Secondly, in order to improve the capability of classifiers, an active sample selection strategy is proposed by optimizing a bag Fisher information function instead of the bag likelihood function, which dynamically chooses most discriminative samples for classifier training. Furthermore, we integrate the co-training criterion into algorithm to update the appearance model accurately and avoid error accumulation. Finally, we evaluate our system on challenging sequences and an indoor environment in a laboratory. And the experiment results demonstrate that the proposed methods can stably and robustly track moving object.

• ETRI Journal
• /
• v.44 no.3
• /
• pp.462-475
• /
• 2022

Although researchers have proposed numerous techniques for speech emotion recognition, its performance remains unsatisfactory in many application scenarios. In this study, we propose a speech emotion recognition model based on a genetic algorithm (GA)-decision tree (DT) fusion of deep and acoustic features. To more comprehensively express speech emotional information, first, frame-level deep and acoustic features are extracted from a speech signal. Next, five kinds of statistic variables of these features are calculated to obtain utterance-level features. The Fisher feature selection criterion is employed to select high-performance features, removing redundant information. In the feature fusion stage, the GA is is used to adaptively search for the best feature fusion weight. Finally, using the fused feature, the proposed speech emotion recognition model based on a DT support vector machine model is realized. Experimental results on the Berlin speech emotion database and the Chinese emotion speech database indicate that the proposed model outperforms an average weight fusion method.

• Journal of Trauma and Injury
• /
• v.24 no.2
• /
• pp.89-94
• /
• 2011

Purpose: A few studies have assessed the factors affecting the prognoses for major trauma patients and those improving the circumstances when dealing with the trauma system. In that light, we analyzed factors, such as pre-hospital factors, the time to admission, the length of stay in the emergency department (ED) and emergency operation, influencing the outcomes for trauma patients. Methods: The patients who visited our emergency department from April 1, 2009, to February 29, 2011, due to major trauma were enrolled in the study. The inclusion criterion was a revised trauma score (RTS) < 7 or injury severity score (ISS) ${\geq}$ 16. We used reviews of medical records, to analyze each step of emergency medical care with respect to patients' sex, age, visit time and visit date. Continuous variables were described as a median with an interquartile range, and we compared the variables between the survival and the mortality groups by using the Mann-Whitney U test. Fisher's exact test was used for nominal variables. Using the variables that showed statistical significance in univariate comparisons, we performed a logistic regression analysis, and we tested the model's adequacy by the using the Hosmer-Lemeshow method. Results: A total of 261 patients with major trauma satisfied either the RTS score criterion or the ISS score criterion. Excluding 12 patients with missing data, 249 patients were included in this study. The overall mortality rate was 16.9%. Time to ED arrival, time to admission, time of ED stay, RTS, ISS, and visit date being a holiday showed statistically significant differences between the survival and the mortality groups in the univariate analysis. RTS, ISS, length of ED stay, and visit date being a holiday showed statistical significance in the multivariate analysis. Conclusion: The mortality rate did not show a significant relationship with the time to ED arrival, use of 119, on time to admission. Rather, it elicited a quite significant correlation with the trauma scoring system (RTS and ISS), the time of ED stay, and the visit date being a holiday.

• Korean Journal of Biological Psychiatry
• /
• v.11 no.2
• /
• pp.165-172
• /
• 2004

Objective:Although polymorphism of dopamine transporter gene(DAT1) has been considered to be implicated in the pathogenesis of social phobia, previous investigations have been inconsistent and controversial. The authors investigated the relationship between DAT1 polymorphism and social phobia in Koreans. Methods:DAT1 and alleles of fifty subjects who met DSM-IV criterion of social phobia, and those of age- & sex- matched fifty normal controls in Korea were compared. Additionally, patients were grouped into generalized(33) and nongeneralized(17) types and DAT1 polymorphism was compared with that of age- & sex- matched controls. DAT1 with variable number of tandem repeats(VNTR) were determined by using polymerase chain reaction. To compare the distribution of the DAT1 polymorphism between different groups, Fisher's exact test was used. Results:There were no significant differences in either genotypic(p=0.451) or allelic(p=0.452) distributions between the social phobia patients and the controls. There also were no differences in genotypic distribution between subtypes of social phobia patients and the controls. Conclusion:We couldn't find any association between DAT1 polymorphism and social phobia. Further studies including larger number of samples and diverse clinical variables should be conducted to elucidate the present findings.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.42 no.4 s.304
• /
• pp.33-42
• /
• 2005

This paper proposes a classification method for gene expression data, using membership function and neural network. The gene expression is a process to produce mRNA and protains which generate a living body, and the gene expression data is important to find out the functions and correlations of genes. Such gene expression data can be obtained from DNA 칩 massively and quickly. However, thousands of gene expression data may not be useful until it is well organized. Therefore a classification method is necessary to find the characteristics of gene data acquired from the gene expression. In the proposed method, a set of gene data is extracted according to the fisher's criterion, because we assume that selected gene data is the well-classified data sample. However, the selected gene data does not guarantee well-classified data sample and we calculate feature values using membership function to reduce the influence of outliers in gene data. Feature vectors estimated from the selected feature values are used to train back propagation neural network. The experimental results show that the clustering performance of the proposed method has been improved compared to other existing methods in various gene expression data.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.5
• /
• pp.183-191
• /
• 2010

Unique and invariant biometric characteristics have been used for secure user authentication. Storing original biometric data is not acceptable due to privacy and security concerns of biometric technology. In order to enhance the security of the biometric data, the cancelable biometrics was introduced. Using revocable and non-invertible transformation, the cancelable biometrics can provide a way of more secure biometric authentication. In this paper, we present a new cancelable bits extraction method for the facial data. For the feature extraction, the Subpattern-based Principle Component Analysis (PCA) is adopted. The Subpattern-based PCA divides a whole image into a set of partitioned subpatterns and extracts principle components from each subpattern area. The feature extracted by using Subpattern-based PCA is discretized with a helper data based method. The elements of the obtained bits are evaluated and ordered according to a measure based on the fisher criterion. Finally, the most discriminative bits are chosen as the biometric bits string and used for authentication of each identity. Even if the generated bits string is compromised, new bits string can be generated simply by changing the helper data. Because, the helper data utilizes partial information of the feature, the proposed method does not reveal privacy sensitive biometric information of the user. For a security evaluation of the proposed method, a scenario in which the helper is compromised by an adversary is also considered.

• Journal of Trauma and Injury
• /
• v.25 no.4
• /
• pp.247-253
• /
• 2012

Purpose: Hemorrhage is a main cause of death in trauma patients. The goal of this study is to describe the characteristics of trauma patients with massive bleeding and to evaluate the prognostic factors concerning their survival. Methods: This study was performed retrospectively and included trauma patients with massive bleeding who had been treated from March 2007 to August 2012. The inclusion criterion was patients who received more than 10 U of packed red blood cells within the first 24 hours after visiting the emergency department. Based on their medical records, we collected data in terms of demographic findings, mechanisms of injury, initial clinical and laboratory findings, methods for hemostasis (emergency surgery and/or angioembolization), transfusion, injury severity score (ISS), revised trauma score (RTS) and trauma and injury severity score (TRISS). We used the Mann-Whitney U test and Fisher's exact test to compare the variables between the patients that survived and those that did not. We performed a logistic regression analysis with the significant variables from the univariate test. Results: Thirty-two(32) patients were enrolled. The main mechanisms of injury were falls and motor vehicle accidents. The mean transfusion amount of packed red blood cells (PRBC) was 17.4 U. The mean elapsed time for the first hemostasis (surgery or embolization) was 3.5 hours. The initial technical success rates were 83.3%(15/18) in angioembolization and 66.7%(8/12) in surgery. The overall mortality rate was 34.4%(11/32). The causes of death were bleeding, brain swelling and multiple organ failure. The ISS(25.5 vs 46.3, p=0.000), TRISS(73.6 vs 45.1, p=0.034) and base excess(<-12 mmol/L, p=0.020) were significantly different between the patients who survived and those who did not. Conclusion: The ISS was a prognostic factor for trauma patients with massive bleeding.