• The Journal of the Acoustical Society of Korea
• /
• v.19 no.2
• /
• pp.40-45
• /
• 2000

In speech corpus generation and speech recognition, it is sometimes needed to segment the input speech data without any prior knowledge. A method to accomplish this kind of segmentation, often called as blind segmentation, or acoustic segmentation, is to find boundaries which minimize the Euclidean distances among the feature vectors of each segments. However, the use of this metric alone is prone to errors because of the fluctuations or variations of the feature vectors within a segment. In this paper, we introduce the principal component analysis method to take the trend of feature vectors into consideration, so that the proposed distance measure be the distance between feature vectors and their projected points on the principal components. The proposed distance measure is applied in the LBDP(level building dynamic programming) algorithm for an experimentation of continuous speech segmentation. The result was rather promising, resulting in 3-6% reduction in deletion rate compared to the pure Euclidean measure.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.13 no.2
• /
• pp.157-163
• /
• 2013

In this paper, we propose a human detection technique using thermal imaging camera. The proposed method is useful at night or rainy weather where a visible light imaging cameras is not able to detect human activities. Under the observation that a human is usually brighter than the background in the thermal images, we estimate the preliminary human regions using the statistical confidence measures in the gray-level, brightness histogram. Afterwards, we applied Gaussian filtering and blob labeling techniques to remove the unwanted noise, and gather the scattered of the pixel distributions and the center of gravities of the blobs. In the final step, we exploit the aspect ratio and the area on the unified object region as well as a number of the principal components extracted from the object region images to determine if the detected object is a human. The experimental results show that the proposed method is effective in environments where visible light cameras are not applicable.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2017.06a
• /
• pp.83-83
• /
• 2017

The National Agrobiodiversity Center (NAS, RDA, Republic of Korea) has continually collected new valuable genetic resources. In this study, we regenerated conserved kidney bean (Phaseolus vulgaris L.) germplasm which couldn't be available because of seed quantity and quality, and we also surveyed their morphological characters for the sustainable utilization. A total of 431 kidney bean accessions were regenerated and 18 morphological traits were surveyed according to the characterization guideline of RDA Genebank. Among the surveyed traits, flowering time ranged from May 23 to September 4 and 73.8% of tested accessions were mainly flowering in June. The maturity time ranged from July 1 to October 15 and main flowering time was July (91.4%). For plant type, 270 accs (62.6%) were climbing type followed by medium type of 86 accs (20.0%) and dwarf type of 65 accs (15.1%). The seed coat colors were various; yellow (34.6%), white (22.3%), brown (17.9%), red (10.7%), black (5.8%), violet (11%), pink (1.4%), navy (0.9%). Principal component analysis indicated that five principal components (PCs) with Eigen values >1 accounted for more than 65.8% variability. The first PC was more related to growth habits such as growth type, flowering time, and plant type. The second and third PCs showed higher values of the pigment characters such as seed coat color, flower color, and pod color. In fourth and fifty PCs, there were the higher positive values of the pod shapes. Our results provided insight into the characteristics kidney beans, thus the utilization basis of kidney beans might be elevated for bio-industry.

• Nuclear Engineering and Technology
• /
• v.48 no.5
• /
• pp.1184-1191
• /
• 2016

This paper introduces the development of a transient monitoring system to detect the early stage of a transient, to identify the type of the transient scenario, and to inform an operator with the remaining time to turbine trip when there is no operator's relevant control. This study focused on the transients originating from a secondary system in nuclear power plants (NPPs), because the secondary system was recognized to be a more dominant factor to make unplanned turbine-generator trips which can ultimately result in reactor trips. In order to make the proposed methodology practical forward, all the transient scenarios registered in a simulator of a 1,000 MWe pressurized water reactor were archived in the transient pattern database. The transient patterns show plant behavior until turbine-generator trip when there is no operator's intervention. Meanwhile, the operating data periodically captured from a plant computer is compared with an individual transient pattern in the database and a highly matched section among the transient patterns enables isolation of the type of transient and prediction of the expected remaining time to trip. The transient pattern database consists of hundreds of variables, so it is difficult to speedily compare patterns and to draw a conclusion in a timely manner. The transient pattern database and the operating data are, therefore, converted into a smaller dimension using the principal component analysis (PCA). This paper describes the process of constructing the transient pattern database, dealing with principal components, and optimizing similarity measures.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.9
• /
• pp.243-251
• /
• 2022

In this paper, we propose a phenomenon that analyze the impact of market sentiment on China's real estate market through the perspective of behavioral economics. Previously, real estate market analyzation basically focus on some fundamental principles which include market price, monetary policies and income, etc. However, little research has explored market sentiment and its influence. By using principal components analysis (PCA), this study first creates buyer's sentiment and seller's sentiment to measure the heat of China's real estate market. Different from using traditional estimation method, the vector autoregressive model (VAR) is used to analyze how both sentiments affect real estate return. The overall results show that from unit root test and impulse response analyzation, the impact of seller's sentiment is positive to real estate market while buyer's sentiment is negative. At the same time, the higher seller's sentiment will have different influence on the housing market compared with the higher buyer's sentiment.

• Bulletin of the Korean Chemical Society
• /
• v.28 no.12
• /
• pp.2389-2395
• /
• 2007

Headspace solid phase trapping solvent extraction (HS-SPTE) and GC-MS was applied for the characterization of volatile flavors from fennel, anise seed, star-anise, dill seed, fennel bean, and Ricard aperitif liquor. Tenax was used for HS-SPTE adsorption material. Recoveries, precision, linear dynamic ranges, and the limit of detection in the analytical method were validated. There were some similarities and distinct differences between fennel-like samples. The Korean and the Chinese fennels contained trans-anethole, (+)-limonene, anisealdehyde, methyl chavicol as major components. The volatile aroma components from star anise were characterised by rich trans-anethole, (+)-limonene, methyl chavicol, and anisaldehyde. Additionally, principal component analysis (PCA) has been used for characterizing or classifying eight different fennel-like samples according to origin or other features. A quite different pattern of dill seed was found due to the presence of apiol (dill).

• Journal of Korea Multimedia Society
• /
• v.15 no.1
• /
• pp.40-50
• /
• 2012

Face recognition is a biometric technology used to identify individuals based on facial feature information. Previous studies of face recognition used features including the eye, mouth and nose; however, there have been few studies on the effects of using other facial components, such as the eyebrows and chin, on recognition performance. We measured the recognition accuracy affected by these facial components, and compared the differences between computer-based and human-based facial recognition methods. This research is novel in the following four ways compared to previous works. First, we measured the effect of components such as the eyebrows and chin. And the accuracy of computer-based face recognition was compared to human-based face recognition according to facial components. Second, for computer-based recognition, facial components were automatically detected using the Adaboost algorithm and active appearance model (AAM), and user authentication was achieved with the face recognition algorithm based on principal component analysis (PCA). Third, we experimentally proved that the number of facial features (when including eyebrows, eye, nose, mouth, and chin) had a greater impact on the accuracy of human-based face recognition, but consistent inclusion of some feature such as chin area had more influence on the accuracy of computer-based face recognition because a computer uses the pixel values of facial images in classifying faces. Fourth, we experimentally proved that the eyebrow feature enhanced the accuracy of computer-based face recognition. However, the problem of occlusion by hair should be solved in order to use the eyebrow feature for face recognition.

• Korean Journal of Plant Resources
• /
• v.19 no.2
• /
• pp.300-307
• /
• 2006

To clarify the morphological and geographical differentiation among the polyploid complexes of L. lancifolium collections in Korea, the mo게hological variation of 173 accessions were analyzed by ANOVA (one-way analysis of variance) and PCA (principal component analysis) on the basis of 38 morphological characters. 173m accessions were grouped into 78 diploids and 95 triploids by ploid levels and the triploids separated into 75 inland triploids (all around the Korea) and 20 island triploids (Backryung-do and Sochung-do, westemmost and northernmost islands of Korea) by geographic distribution and morphology. Island triploids showed significant morphological differences with inland triploids in ANOVA by many floral and leaf characters. In PCAs, diploids were separated from inland triploids by having longer plant height, smaller flower characters, higher pollen fertility and more stomata. The first four principal components accounted for 44.1% of the total variation. Plots of the island and inland groups for the first and second principal components separated each other with slight overlapping. Although the ploid forms are different between diploid and island triploid, island triploids were more closely overlapped with diploids by principal component 1 and 2 than inland triploids. This reflects that the whole external morphology of island triploids are similar to that of diploids. This, the phenotypic differentiation between inland and island triploids seems to be partly related to their geographical origins.

• Mass Spectrometry Letters
• /
• v.4 no.4
• /
• pp.59-66
• /
• 2013

A metabolomics study was conducted to identify urinary biomarkers for breast cancer, using gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS), analyzed by principal components analysis (PCA) as well as a partial least squares-discriminant analysis (PLS-DA) for a metabolic pattern analysis. To find potential biomarkers, urine samples were collected from before- and after-mastectomy of breast cancer patients and healthy controls. Androgens, corticoids, estrogens, nucleosides, and polyols were quantitatively measured and urinary metabolic profiles were constructed through PCA and PLS-DA. The possible biomarkers were discriminated from quantified targeted metabolites with a metabolic pattern analysis and subsequent screening. We identified two biomarkers for breast cancer in urine, ${\beta}$-cortol and 5-methyl-2-deoxycytidine, which were categorized at significant levels in a student t-test (p-value < 0.05). The concentrations of these metabolites in breast cancer patients significantly increased relative to those of controls and patients after mastectomy. Biomarkers identified in this study were highly related to metabolites causing oxidative DNA damage in the endogenous metabolism. These biomarkers are not only useful for diagnostics and patient stratification but can be mapped on a biochemical chart to identify the corresponding enzyme for target identification via metabolomics.

• Resources Recycling
• /
• v.26 no.1
• /
• pp.69-77
• /
• 2017

Used small household appliances(small e-waste) consists of a variety of complex materials and components. The small e-waste is mainly composed of plastics and an important potential source of waste plastic. The black plastics, particularly are very difficult to separate by resin type and therefore these are mainly recycled in the form of a mixtures. In the present study, the sorting technologies such as gravity and electro static separation, near-infrared ray(NIR) and IR/Raman optical sorting separation on mixture of black plastics were analyzed and their limitations on sorting process were also investigated. The Laser Induced Breakdown Spectroscopy(LIBS) spectrum of each black plastics was used for identification of black plastics by resin type, and after analyzing the normalization operation, Principal Component Analysis(PCA) was carried out. The spectrum data was optimized through PCA process. In order to improve the identification accuracy and sorting efficiency of black plastics, it is necessary to design a classifier with high efficiency and to improve the performance and reliability of the classifier by applying the field of intelligent algorithms.