• The Journal of the Acoustical Society of Korea
• /
• v.24 no.1
• /
• pp.58-68
• /
• 2005

In this paper, we propose a scheme to improve discriminative property in the feature compensation method for robust speech recognition under noisy environments. The estimation of noisy speech model used in existing feature compensation methods do not guarantee the computation of posterior probabilities which discriminate reliably among the Gaussian components. Estimation of Posterior probabilities is a crucial step in determining the discriminative factor of the Gaussian models, which in turn determines the intelligibility of the restored speech signals. The proposed scheme employs minimum classification error (MCE) training for estimating the parameters of the noisy speech model. For applying the MCE training, we propose to identify and determine the 'competing components' that are expected to affect the discriminative ability. The proposed method is applied to feature compensation based on parallel combined mixture model (PCMM). The performance is examined over Aurora 2.0 database and over the speech recorded inside a car during real driving conditions. The experimental results show improved recognition performance in both simulated environments and real-life conditions. The result verifies the effectiveness of the proposed scheme for increasing the performance of robust speech recognition systems.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.22 no.5
• /
• pp.656-662
• /
• 2012

In this paper, we describe a new summarizing method based on a graph-based and a sense-based analysis. In the graph-based analysis, we convert sentences in a document into word vectors and calculate the similarity between each sentence using LSA. We reflect this similarity of sentences and the rarity scores of words in sentences to define weights of edges in the graph. Meanwhile, in the sense-based analysis, in order to determine the sense of words, subjectivity or objectivity, we built a database which is extended from the golden standards using Wordnet. We calculate the subjectivity of sentences from the sense of words, and select more subjective sentences. Lastly, we combine the results of these two methods. We evaluate the performance of the proposed method using classification games, which are usually used to measure the performances of summarization methods. We compare our method with the MS-Word auto-summarization, and verify the effectiveness of ours.

• Journal of Hydrogen and New Energy
• /
• v.23 no.3
• /
• pp.274-284
• /
• 2012

Recently, gasification technology for coal blended with biomass has been an issue. Especially, An advantages of coal blended with biomass are 1) obtaining high cold gas efficiency, 2) obtaining syn-gas of high-high heating value (HHV), and 3) controlling occurrence of $CO_2$. In this study, the efficiency and characteristic of 300 MW Shell type gasifier were predicted using CFD simulation. The CFD simulation was performed for biomass coal blending ratios of 0~0.2, 0.5, 1 and $O_2$/fuel ratios of 0.5~0.84. Kinetic parameters (A, $E_a$) obtained by $CO_2$ gasification experiment were used as inputs for the simulation. In results of CFD simulation, residence times of particle in 300MW Shell type gasifer presented as 7.39 sec ~ 13.65 sec. Temperature of exit increased with $O_2$/fuel ratio as 1400 K ~ 2800 K, while there is not an effects of biomass coal blending ratios. Considering both aspects of temperature for causing wall slagging and high cold gas efficiency, the optimal $O_2$/fuel ratio and blending ratio were found to be 0.585 and 0.05, respectively.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.34-36
• /
• 2022

In the existing text sentiment analysis models, the entire text is usually directly modeled as a whole, and the hierarchical relationship between text contents is less considered. However, in the practice of sentiment analysis, many texts are mixed with multiple emotions. If the semantic modeling of the whole is directly performed, it may increase the difficulty of the sentiment analysis model to judge the sentiment, making the model difficult to apply to the classification of mixed-sentiment sentences. Therefore, this paper proposes a sentiment analysis model BHGCN that considers the text hierarchy. In this model, the output of hidden states of each layer of BERT is used as a node, and a directed connection is made between the upper and lower layers to construct a graph network with a semantic hierarchy. The model not only pays attention to layer-by-layer semantics, but also pays attention to hierarchical relationships. Suitable for handling mixed sentiment classification tasks. The comparative experimental results show that the BHGCN model exhibits obvious competitive advantages.

• Membrane Journal
• /
• v.25 no.5
• /
• pp.373-384
• /
• 2015

In the past few decades, polymeric membrane has played an important role in gas separation applications. For the separation of $CO_2$, one of greenhouse gases, high permselectivity, long-term stability and scale-up are needed. However, conventional polymeric membranes have shown a trade-off relation between permeability and selectivity while inorganic materials are highly permeable but expensive. Mixed matrix membranes (MMMs) combining the advantages of both polymeric and inorganic materials have become a possible breakthrough for the next-generation gas separation membranes. The MMMs could be either symmetric or asymmetric but the latter is more preferred to improve the permeance. Important factors influencing the MMM fabrication include homogeneous distribution of inorganic particles and good interfacial contact between inorganic filler and organic matrix. Recently, metal organic frameworks (MOFs) have received much attention as a new class of porous crystalline materials and a potential candidate for $CO_2$ separation. Zeolitic imidazolate frameworks (ZIFs), a sub-branch of MOFs, are the most widely used in MMMs due to small particle size and appropriate pore size for $CO_2$ separation. One of the major issues associated with the incorporation of porous particles in a polymeric membrane is to control the microstructure of the porous particle materials such as particle size, orientation, and boundary conditions etc. In this review, major challenges surrounding MMMs and the strategies to tackle these challenges are given in detail.

• Korean Journal of Environment and Ecology
• /
• v.27 no.3
• /
• pp.280-304
• /
• 2013

This study was carried out to investigate the flora of Mt. Gilsangsan (Ganghwa-Isl.). The vascular plants were collected 14 times (from March to November 2012), and were identified as 595 taxa; 122 families, 359 genera, 529 species, 7 subspecies, 51 varieties, 2 forms and 3 hybrids. The plant formation of Mt. Gisangsan is the deciduous broad-leaved and conifer mixed forest which is the common one in the middle part of Korean peninsula. All most mountain covered with young secondary forest which is mainly composed of Pinus and Quercus. The family represented by the largest number of species diversity were Asteraceae (11.%), Poaceae (10.7%), Cyperaceae (6.38%), Fabaceae (3.86%) and Rosaceae (3.52%) respectively. Korean endemic plants of this area were 12 taxa; Asarum misandrum, Clematis brachyura, Philadelphus schrenkii, Forsythia koreana, Polygonatum infundiflorum, etc. The red list of vascular plants according to IUCN valuation basis were examined, endangered species (EN) Abies koreana, Near Threatened species (NT) both Wikstroemia trichotoma and Sparganium erectum, Least Concern species (LC) Platycladus orientalis, Swertia diluta var. tosaensis and Iris minutoaurea and by Not Evaluate species (NE) Polygonatum infundiflorum. The floristic regional indicator plants found in this area were 45 taxa comprising two taxa of grade V and VI, nine taxa of grade III, three taxa of grade II, 29 taxa of gradeI. The naturalized plants were identified as 64 taxa and the percent of naturalized index (NI) was 10.7% of total 595 taxa vascular plants. Ecosystem disturbance wild plants were consisted of eight taxa; Rumex acetosella, Sicyos angulatus, Solanum carolinense, Ambrosia artemisiifolia, Ambrosia trifida, Aster pilosus, Hypochaeris radicata and Lactuca scariola. Furthermore, therophytes (26.5%), hemicryptophytes (26.0%), Geophyte (14.0%) showed high proportional ratio in life form spectrum.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.5
• /
• pp.7-13
• /
• 2020

This paper focuses on detecting abnormal patterns of respiration of humans. In this study, a contact-based device was used to acquire both normal and abnormal respiration signals. To this end, this paper reports the development of a monitoring system to investigate the respiratory status of humans in a normal environment. This work aims to classify the respiratory status, i.e., normal and abnormal status, quantitatively. The respiration signal is acquired using a contact-based medical device (BIOBPAC), and noise reduction is carried out before classifying the respiratory status. To reduce noise, a mixed filter that combines the Savitzky-Golay filter and Median filter is applied to the acquired respiration signals. The inter-class distance is maximized, and the intra-class distance is minimized. The proposed algorithm is straightforward and can be applied to a practical environment. In addition, the experimental results are provided to substantiate the proposed approach.

• Journal of the Korean Data and Information Science Society
• /
• v.21 no.4
• /
• pp.689-697
• /
• 2010

In this study, we propose a new procedure for improving the character recognition of text area extracted from video images. The recognition of strings extracted from video, which are mixed with Hangul, English, numbers and special characters, etc., is more difficult than general character recognition because of various fonts and size, graphic forms of letters tilted image, disconnection, miscellaneous videos, tangency, characters of low definition, etc. We improved the recognition rate by taking commonly used letters and leaving out the barely used ones instead of recognizing all of the letters, and then using SVM and RDA character recognition methods. Our numerical results indicate that combining SVM and RDA performs better than other methods.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.6
• /
• pp.1159-1175
• /
• 2013

The lift table summarized and calculated the life expectancy at each ages according to aging. In this study, we tried to re-classify and summarized the application of the life table because the life table can be applied to several research and industry area. We utilized the whole papers published in Korea and international until 2011 and we considered several classification standards based on application, base period, gender, and observation period. Each standard divides the life tables into two or three categories. The standard of application groups them into general and applied life tables. The standard of base period is divided into two parts: abridge life table using the unit of 5 years and complete life table of one year. According to gender, life tables are classified in male, female, and unisex life tables. According to complete life table, they are period life tables and cohort life tables. This study contributes to inform how life tables can be employed in many other areas by analyzing how the life tables are constructed.

• Journal of Korean Society of Forest Science
• /
• v.104 no.3
• /
• pp.368-374
• /
• 2015

On the basis of vegetation data by point quarter sampling method, the natural forest of Jirisan(Mt.) was classified into eight forest types by cluster analysis. They were Quercus mogolica forest type, Fraxinus mandshurica - Betula costata forest type, Mixed mesophytic forest type, Abies koreana forest type, Carpinus laxiflora forest type, Quercus serrata forest type, Pinus densiflora forest type, and Quercus variabilis forest type. Then, succession index was estimated for each forest type so as to evaluate succession process comparatively among forest types. The results showed that Carpinus laxiflora forest type had highest succession index of 219.7, followed by Mixed mesophytic forest type with little difference of the index of 218.3. Pinus densiflora forest type had lowest index. Succession indices were hardly correlated with species diversity indices of forest types. We presumed that the higher value of succession index a forest type had, the closer toward the climax forest. However, the estimated index was not supposed to be absolute level of successional stage, but it could play a role of comparative assessment in the position of the seral stage among forest types.