• MALSORI
• /
• no.63
• /
• pp.125-137
• /
• 2007

Recently, many techniques have been proposed to improve the noise robustness for speaker verification. In this paper, we consider the feature recombination technique in multi-band approach. In the conventional feature recombination for speaker verification, to compute the likelihoods of speaker models or universal background model, whole feature components are used. This computation method is not effective in a view point of multi-band approach. To deal with non-effectiveness of the conventional feature recombination technique, we introduce a subband likelihood computation, and propose a modified feature recombination using subband likelihoods. In decision step of speaker verification system in noise environments, a few very low likelihood scores of a speaker model or universal background model cause speaker verification system to make wrong decision. To overcome this problem, a reliable feature selection method is proposed. The low likelihood scores of unreliable feature are substituted by likelihood scores of the adaptive noise model. In here, this adaptive noise model is estimated by maximum a posteriori adaptation technique using noise features directly obtained from noisy test speech. The proposed method using subband-based reliable feature selection obtains better performance than conventional feature recombination system. The error reduction rate is more than 31 % compared with the feature recombination-based speaker verification system.

• The KIPS Transactions:PartB
• /
• v.15B no.3
• /
• pp.229-236
• /
• 2008

This paper examines effectiveness of using color images instead of grayscale ones for automatic lipreading. First, we show the effect of color information for performance of humans' lipreading. Then, we compare the performance of automatic lipreading using features obtained by applying principal component analysis to grayscale and color images. From the experiments for various color representations, it is shown that color information is useful for improving performance of automatic lipreading; the best performance is obtained by using the RGB color components, where the average relative error reductions for clean and noisy conditions are 4.7% and 13.0%, respectively.

• Journal of English Language & Literature
• /
• v.53 no.2
• /
• pp.335-357
• /
• 2007

This study aims to investigate the pragmatic features of Korean EFL learners' interlanguage in the communicative act of complaining. Since a complaint, by its nature, is likely to cause offence, thereby threatening the social relationship between the speaker and the hearer, making a complaint in a polite manner is of crucial importance in maintaining harmonious social relationships. However, very little research has been carried out on the complaint speech act performances of Korean EFL learners. In particular, studies which attempt to examine the effect of social status on the choice of complaint speech act strategies are rare. The present study compared the complaint speech act performances of Korean EFL learners and those of native speakers of English with respect to social status. 24 Korean EFL learners and 28 native speakers of English participated and the data were collected via a Discourse Completion Test. The findings revealed that Korean EFL learners differed from native English speakers in the use of complaint strategies. These results indicate that Korean EFL learners lack certain important skills necessary to make complaints appropriately, suggesting the need for the foreign language learners to develop a more extensive pragmatic knowledge of complaint strategies.

• Sijohaknonchong
• /
• v.20
• /
• pp.113-139
• /
• 2004

This article intends to illuminate how the men. leading agents in Saseol-sijo - musical performers. writers of lyrics, patrons. composers. compilers of Sijo anthologies, audience. etc. - In the Late Choson period, viewed or recognized women and how their understanding of women was reflected in the texts. Working with texts with the theme of 'Love,' this article starts with categorizing two types of love: the first type, 'lovelorn heart' focusing on unilateral pining for a single lover who is absent now and the second type. 'physical love' concentrating on bilateral sexual intercourse. In addition to the types of love, the gender of poetic speakers, distinct from real poets is vital to characterize the discourse of love. According to these two factors. texts in question fall into four groups: texts that a female speaker displays her lovelorn heart('Type 1'), those where she speaks about her sexual experiences('Type 2'), those where a male speaker sings his lovelorn heart('Type 3'), and those where he describes his sexual experiences('Type 4'). Of these. 'Type 2' and 'Type 3' are key to understanding of the men's view of women. With respect to the configuration of the theme of 'Love,' it should be noted that in Korean literary history, the nim or a 'sweetheart' had signified the totality of value or a perfect entity which makes one's life meaningful and that 'Type 1,' the pattern that a female subject expresses her love toward male min, had constituted a traditional way to convey the theme of 'Love.' In terms of this connotation of min. a remarkable increase of 'Type 3' implying the increase of male speakers, reveals the extent to which women, the male speakers' min, accomplished their entry into a 'sacred area' -the position of mm-in which only men had occupied; females are focused and centralized. This article considers this phenomenon as an exhibition of the upgrade of women's significance and weight in the Late Choson society and as an index of 'modernity.' Meanwhile, given that most of the Saseol-sijo poets are men, the emergence of the 'Type 2' texts in which male poets have female speakers disclose their sexual experiences, demonstrates a representative example that women are degraded to be a means of men's pleasure; for this situation gives men more pleasure than when male speakers reveal their sexual experiences. Not only 'Type 2,' but texts group which basically belongs to 'Type I' and conveys the theme of 'Loyalty' through the female voice by substituting rulers-subjects relation for men-women relation, also falls under the same case. For men employ female voice as a poetic device in order to stress the theme of 'Loyalty' This article regards this phenomenon as an index of 'pre-modernity,' in the sense that in a pre-modem society, specifically in Early Choson, male-oriented value system dominates, thereby alienating women. As it is well known, the Late Choson is marked by a transitional period from a pre-modem society to a modem society. Therefore the ambivalence of the premodern and the modem can be found mixed in every segment of the society. The dual aspects of the masculine view of women in Saseol-sijo constitutes one example. The significance of the Saseol-sijo in Korean literary history can be found in this phenomenon.

• (The)Study of the Eastern Classic
• /
• no.71
• /
• pp.231-258
• /
• 2018

This paper examines the differences of two different versions of "Cheonuisogameonhae" in terms of stylistics and investigates factors affecting the differences. The interpretations between the woodblock edition and the manuscript edition might be different depending on assumed range of readership, and the stylistic differences between two editions might be different depending on the possibility of extension of the reading population. Thus, this paper examines how stylistic effects are reflected in inter-relations between a translator as a speaker and readers as listeners according to speaker intentions. In Chapter 2, the stylistic differences reflected from two difference editions are examined in terms of the expression of a writer's respect, emotions, and formal consciousness to readers. The expressions of a writer's respect are more clearly emerged in the manuscript edition than in the woodblock edition. The honorific expression of a subject, '-gyeo?dsyeo', and the honorific expression of a writer, '-s?p-', are more frequently used in the manuscript edition than in the woodblock edition. In order to express positive emotions, exclamation endings are used in the manuscript edition, which shows the writer's strong emotional sympathy with readers' words and behaviors. On the other hand, in the woodblock edition, '-이' is used after names in order to treat rebellious subjects and people involved in conspiracy contemptuously by the use of informal forms. In addition, affirmative sentences in the manuscript edition and double negative sentences in the woodblock edition are used respectively, which intends to strongly emphasize a king's will and the appropriateness of the will. The writer's formal consciousness to readers are found in the way of writing names of people and places in Korean. Chinese characters are generally used two show formal consciousness; thus, names of people and places are expressed in Chinese characters in the woodblock edition. In Chapter 3, factors that made the stylistic differences between two editions are examined. The factors causing stylistic differences are examined in terms of the purpose of the interpretation, the class and range of the reading population, a writer's attitudes toward readers, and the face-to-fact situation of a writer and readers.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.8
• /
• pp.96-105
• /
• 1998

The aim of this paper is to propose the method of WDHMM(Weighted DHMM), using the MLP-VQ for the improvement of speaker-independent connect-digit recognition system. MLP neural-network output distribution shows a probability distribution that presents the degree of similarity between each pattern by the non-linear mapping among the input patterns and learning patterns. MLP-VQ is proposed in this paper. It generates codewords by using the output node index which can reach the highest level within MLP neural-network output distribution. Different from the old VQ, the true characteristics of this new MLP-VQ lie in that the degree of similarity between present input patterns and each learned class pattern could be reflected for the recognition model. WDHMM is also proposed. It can use the MLP neural-network output distribution as the way of weighing the symbol generation probability of DHMMs. This newly-suggested method could shorten the time of HMM parameter estimation and recognition. The reason is that it is not necessary to regard symbol generation probability as multi-dimensional normal distribution, as opposed to the old SCHMM. This could also improve the recognition ability by 14.7% higher than DHMM, owing to the increase of small caculation amount. Because it can reflect phone class relations to the recognition model. The result of my research shows that speaker-independent connected-digit recognition, using MLP-VQ and WDHMM, is 84.22%.

• (The)Study of the Eastern Classic
• /
• no.41
• /
• pp.121-148
• /
• 2010

The aim of this paper is to look into the characteristics and its value of Jo Uin(曺友仁)'s Jadosa(自悼詞). I compared Jadosa with other Yeongungasa(戀君歌辭) works for this purposes in the sides of aspect of Yeongunuisik(戀君意識) and emotionalism. Therefore Jadosa is equipped with space setting called the heavenly world and the earthly world, and has characteristics that a speaker of the earthly world misses a lover of the heavenly world. Also, Jadosa is similar to Samiingok(思美人曲) and Sokmiingok(續美人曲) of Jeong Cheol(鄭澈) because the former borrowed a few phrases and motifs from the latter. However, if I look into Jadosa in greater detail in the sides of emotion or attitude of a speaker, the speaker of Jadosa shows a reproachful attitude Unlike works of Jeong Cheol. And the speaker of Jadosa urges the lover to be aware of his illusion. Finally these differences occur as a political standing and a relation with the king is different in every writer. Accordingly This paper is very worthwhile through comparison of Jadosa and other Yeongungasa works given that I reviewed characteristics and a meaning of Jadosa.

• Journal of the Korean Society for Library and Information Science
• /
• v.52 no.3
• /
• pp.289-334
• /
• 2018

Even though information in many languages other than English is quickly increasing, English is still playing the role of the lingua franca and being accounted for the largest proportion on the web. Therefore, it is necessary to investigate the key features and differences between "information searching behavior using mother tongue as a search language" and "information searching behavior using English as a search language" of users who are non-mother tongue speakers of English to acquire more diverse and abundant information. This study conducted the experiment on the web searching which is applied in concurrent think-aloud method to examine the information searching behavior and the cognitive process in Korean search and English search through the twenty-four undergraduate students at a private university in South Korea. Based on the qualitative data, this study applied the frequency analysis to web search pattern under search language. As a result, it is active, aggressive and independent information searching behavior in Korean search, while information searching behavior in English search is passive, submissive and dependent. In Korean search, the main features are the query formulation by extract and combine the terms from various sources such as users, tasks and system, the search range adjustment in diverse level, the smooth filtering of the item selection in search engine results pages, the exploration and comparison of many items and the browsing of the overall contents of web pages. Whereas, in English search, the main features are the query formulation by the terms principally extracted from task, the search range adjustment in limitative level, the item selection by rely on the relevance between the items such as categories or links, the repetitive exploring on same item, the browsing of partial contents of web pages and the frequent use of language support tools like dictionaries or translators.

• Journal of Science Education
• /
• v.41 no.3
• /
• pp.462-479
• /
• 2017

This study classified the visualization's materials presented in two types of middle school science textbooks to data visualization and infographics, thereafter characteristics of infographics according to science content's domains were analyzed using an analytical framework for infopraphics. The results of this study are as follows : First, visualization materials in middle school science textbooks had more data visualization (23.2%) than infographics (76.8%). Second, there were many differences in information content's category among the science subject's domains. The relationship type of information content in the domain of energy and matter, the function type in life domain, and the location type in the earth domain were prominent. In visual representation's category, the illustration type and comparative analysis type were the most dominant in the all science domains. The horizontal type in layout category occupied the largest proportion in all science domains. Therefore, it is necessary to present the visualization materials with the infographics conveying many informations than the data visualization in middle school science textbooks. And the infographics need to be developed and distributed in various formats in each science domain.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.16 no.2
• /
• pp.196-205
• /
• 1991

This paper is a study on the isolated word recognition of speaker independent which proposes to newly improved MSVQ(multisection vector quantization) recognition system which improve the classical MSVQ recognition system. It is a difference that test pattern has on more section than reference pattern in recognition system 146 DDD area names are selected as recognition vocabulary. 12th LPC cepstral coefficients is used as feature parameter. and when codebook is generated, MINSUM and MINMAX are used in finding the centroid. According to the experiment result. it is proved that this method is better than VQ(vector quantization) recognition methods, DTW(dynamic time warping) pattern matching methods and classical MSVQ methods for recognition rate and recognition time.