• The Journal of the Acoustical Society of Korea
• /
• v.16 no.2
• /
• pp.49-57
• /
• 1997

In this paper, we propose the stochastic pronunciation lexicon model for large vocabulary continuous speech recognition system. We can regard stochastic lexicon as HMM. This HMM is a stochastic finite state automata consisting of a Markov chain of subword states and each subword state in the baseform has a probability distribution of subword units. In this method, an acoustic representation of a word can be derived automatically from sample sentence utterances and subword unit models. Additionally, the stochastic lexicon is further optimized to the subword model and recognizer. From the experimental result on 3000 word continuous speech recognition, the proposed method reduces word error rate by 23.6% and sentence error rate by 10% compare to methods based on standard phonetic representations of words.

• Archives of design research
• /
• v.19 no.4 s.66
• /
• pp.51-60
• /
• 2006

Nowadays, reading text from screens is prevailing in everyday life. The advent of mobile information devices such as a cellular phone, PDA, and e-book reader facilitates us to enjoy various text-based contents any time and anywhere. Most studies comparing screen and paper readability show that screens are less readable than paper. Furthermore, the decrease of line length and number of lines that can be displayed on the screen of mobile information devices deteriorate text readability. This study investigated parameters affecting text readability on small screens and designed a new text layout to improve readability. We suggested a diagonally splitted layout of rectangular column, which is supposed to facilitate eye movement to trace text flow with ease. The experiment comparing readability between a traditional rectangular column and a diagonally splitted column was conducted. The result of experiment revealed that there is no significant difference between the two text layouts in terms of subjective satisfaction of reading task and a level of comprehension. However, in the screen size of $4000mm^2\;and\;8000mm^2$, reading speed was increased 18.9% and 34.0% respectively from a traditional rectangular column to a diagonally splitted column. We conducted a consecutive experiment to scrutinize the cause that improved the performance in readability task remarkably. The readability of text in a traditional rectangular column was compared with a left triangular column and a right triangular column in the condition of $4000mm^2/3:1$ ratio screen. The performance measurements revealed that participants read 21.1% and 67.6% faster respectively with the left triangular column and right triangular column than with the rectangular column. In consequence, the improvement of readability in the diagonally splitted column was attributed mainly to the increase of reading speed in the right triangular column. This research verified that the diagonally splitted column improve text readability on a small screen and this result is expected to make a contribution to designing an efficient text layout for mobile information devices

• Journal of Oral Medicine and Pain
• /
• v.31 no.1
• /
• pp.59-67
• /
• 2006

There exist patients complaining speech problem due to dysesthesia or anesthesia following dental surgical procedure accompanied by local anesthesia in clinical setting. However, it is not clear whether sensory problems in orofacial region may have an influence on motor speech abilities. The purpose of this study was to investigate whether transitory sensory impairment of mandibular nerve by local anesthesia may influence on the motor speech abilities and thus to evaluate possibility of distorted motor speech abilities due to dysesthesia of mandibular nerve. The subjects in this study consisted of 7 men and 3 women, whose right inferior alveolar nerve, lingual nerve and long buccal nerve was anesthetized by 1.8 mL lidocaine containing 1:100,000 epinephrine. All the subjects were instructed to self estimate degree of anesthesia on the affected region and speech discomfort with VAS before anesthesia, 30 seconds, 30, 60, 90, 120 and 150 minutes after anesthesia. In order to evaluate speech problems objectively, the words and sentences suggested to be read for testing speech speed, diadochokinetic rate, intonation, tremor and articulation were recorded according to the time and evaluated using a Computerized Speech $Lab^{(R)}$. Articulation was evaluated by a speech language clinician. The results of this study indicated that subjective discomfort of speech and depth of anesthesia was increased with time until 60 minutes after anesthesia and then decreased. Degree of subjective speech discomfort was correlated with depth of anesthesia self estimated by each subject. On the while, there was no significant difference in objective assessment item including speech speed, diadochokinetic rate, intonation and tremor. There was no change in articulation related with anesthesia. Based on the results of this study, it is not thought that sensory impairment of unilateral mandibular nerve deteriorates motor speech abilities in spite of individual's complaint of speech discomfort.

• Phonetics and Speech Sciences
• /
• v.13 no.3
• /
• pp.79-90
• /
• 2021

This single-subject study examined the effects of repetitive speech training at the word and sentence levels using speech cues on the percentage of correct consonants (PCC) and speech intelligibility of children with cerebral palsy (CP). Three children aged between 5-8 years with a history of CP participated in the study. Thirty-minute intervention sessions were provided four times a week for four weeks. The intervention included repeated training of words and sentences containing target phonemes using two instructions of speech cues, "big mouse" and "strong voice". First, the children improved their average PCC and speech intelligibility, but an effect size analysis indicated that the effect was different for each child, and the effect size for speech intelligibility was higher than for PCC. Second, the intervention effect was generalized to untrained words and sentences. Third, the maintenance effects of PCC and speech intelligibility were very high. These findings suggests that repeated speech training using speech cues is an intervention technique that can help improve PCC and speech intelligibility in children with CP.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.1
• /
• pp.108-113
• /
• 2022

Speech recognition in noise situation is an important skill for effective communication. Hearing In Noise Test (HINT) has been suggested as a clinical tool to evaluate these aspects. However, this tool has not been used widely in domestic clinics. In this study, psychophysical aspects of HINT and burdens in clinical application were analyzed to improve the applicability of the tool. The difficulty in understanding speech in the elderly population is due to hearing loss based on aging of peripheral and central auditory pathways. As typical clinical cases, HINT scores for young and elderly listeners (20s vs 70s) were compared. Four conditions of HINT test were Quiet (Q), Noise Front (NF), Noise Right (NR), and Noise Left (NL). Quantitative scores showed that the elderly listener required more Signal to Noris Ratio (SNR) values than the younger counterpart in noisy situations. Although both showed Binaural Masking Level Difference (BMLD) effect, the strength was smaller in the elder. However, the age-matched normalized data were not established in detail for clinical application. Confirmed usefulness of HINT and the related improvement in clinical measuring procedure were suggested.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2005.05b
• /
• pp.1455-1459
• /
• 2005

K-WEAP(Korea-Water Evaluation And Planning System)은 유역의 물이용 순환체계를 컴퓨터 프로그램으로 구현하고, 수량, 수질, 환경, 수요관리 등을 종합적으로 고려하여 통합수자원계획 수립을 지원하는 전문 모형으로서, 과학기술부와 건설교통부가 공동으로 지원하는 21세기 프론티어 사업인 수자원의 지속적 확보기술사업단의 연구비 지원에 의하여 SEI-B(Stockholm Environment Institute-Boston Center)와 한국건설기술연구원이 공동으로 개발한 모형이다. K-WEAP의 대부분 기능은 기존의 SEI-B가 개발한 WEAP(Water Evaluation And Planning System)에 기반을 두고 있지만, 월 단위 물수지 분석뿐만 아니라 5일 및 임의 시간 단위 물수지분석이 가능하고 물공급 안전도평가와 하천수질모의가 가능하다는 점에서 기존의 WEAP과는 다르며 메뉴와 도움말이 모두 한글로 작성되어있어 국내 사용자들이 이용하기 용이하다. K-WEAP의 기능은 단계적으로 보완 및 개선이 이루어지고 있으며, 현재는 1단계 개발이 끝난 후 2단계 기능개선 작업을 수행하고 있다. 2단계에서 개선하게 될 주요부분은 물수지모형의 개선과 하천수질모형의 개선, 편익산정모형의 개발, 의사결정지원 기능을 고려하는 사용자인터페이스 개선 등이 포함되어 있으며, 2단계 1차 년도에서는 물수지모형과 하천수질모형의 부분적인 개선과 함께 의사결정지원 기능을 고려하는 사용자 인터페이스의 부분적인 개선을 시도하였다. 물수지모형의 개선에서는 하수처리장의 회귀수를 수요처에서 직접 이용할 수 있도록 하였으며, 하천수질모형의 개선부분에서는 기온과 풍속 등의 기후자료를 이용한 수온 모의모형을 개발하였다. 또한, 사용자 인터페이스 부분에서는 사용자의 의사결정을 지원하기 위해 하천유량과 수질 등에 대한 초과비율그래프 조회 기능과 결과를 지도상에서 확인할 수 있는 지도보기기능, 사용자가 필요한 자료를 요약하여 조회할 수 있는 사용자 정의보고서 작성기능을 추가하였다. 개선된 기능을 통해 사용자는 보다 편리한 환경에서 모형을 구동하고 구동결과를 평가 할 수 있을 것으로 기대된다.

• Phonetics and Speech Sciences
• /
• v.11 no.4
• /
• pp.19-33
• /
• 2019

This paper examines the perception of Korean phonemes by Egyptian learners of Korean and presents the learnability gradient of Korean consonants and vowels through High Variability Phonetic Training (HVPT). 50 Egyptian learners of Korean (27 low proficiency learners and 23 high proficiency learners) participated in 10 sessions of HVPT for Korean vowels, word initial and final consonants. Participants were tested on their identification ability of Korean vowels, word initial consonants, and syllable codas before and after the training. The results showed that both low and high proficiency groups did benefit from the training. Low proficiency learners showed a higher improvement rate than high proficiency learners. Based on the HVPT results, a learnability gradient was established to give insights into priorities in teaching Korean sounds to Egyptian learners.

• Journal of Educational Research in Mathematics
• /
• v.24 no.2
• /
• pp.165-180
• /
• 2014

This study analyzed in what ways the instructional materials have been dealing with the fraction as quotient, since the seventh national mathematics curriculum. An analysis of this study urged us to re-consider the content related to the fraction as quotient. First, the fraction as quotient has weakened in the current mathematics textbooks and workbooks in comparison to those developed under the previous curriculum. Second, the contexts of whole number division taught in grades 3 and 4 were not naturally connected to those of the fraction as quotient taught in grade 5. Third, the types of word problems, visual models, and partitioning strategies in the textbooks and the workbooks were partial, and the process of formalization was limited. Building on these results, this study is expected to suggest specific implications which may be taken into account in developing new instructional materials in process.

• Journal of the Korea Convergence Society
• /
• v.12 no.3
• /
• pp.1-7
• /
• 2021

Since Neural Machine Translation (NMT) uses only a limited number of words, there is a possibility that words that are not registered in the dictionary will be entered as input. The proposed method to alleviate this Out of Vocabulary (OOV) problem is Subword Tokenization, which is a methodology for constructing words by dividing sentences into subword units smaller than words. In this paper, we deal with general subword tokenization algorithms. Furthermore, in order to create a vocabulary that can handle the infinite conjugation of Korean adjectives and verbs, we propose a new methodology for subword tokenization training by separating the Jongsung(coda) from Korean syllables (consisting of Chosung-onset, Jungsung-neucleus and Jongsung-coda). As a result of the experiment, the methodology proposed in this paper outperforms the existing subword tokenization methodology.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.8
• /
• pp.325-330
• /
• 2022

This paper conducts the collection and experiment of datasets for deep learning research on sign language such as sign language recognition, sign language translation, and sign language segmentation for Korean sign language. There exist difficulties for deep learning research of sign language. First, it is difficult to recognize sign languages since they contain multiple modalities including hand movements, hand directions, and facial expressions. Second, it is the absence of training data to conduct deep learning research. Currently, KETI dataset is the only known dataset for Korean sign language for deep learning. Sign language datasets for deep learning research are classified into two categories: Isolated sign language and Continuous sign language. Although several foreign sign language datasets have been collected over time. they are also insufficient for deep learning research of sign language. Therefore, we attempted to collect a large-scale Korean sign language dataset and evaluate it using a baseline model named TSPNet which has the performance of SOTA in the field of sign language translation. The collected dataset consists of a total of 11,402 image and text. Our experimental result with the baseline model using the dataset shows BLEU-4 score 3.63, which would be used as a basic performance of a baseline model for Korean sign language dataset. We hope that our experience of collecting Korean sign language dataset helps facilitate further research directions on Korean sign language.