• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.07a
• /
• pp.85-88
• /
• 2011

This paper describes a single channel speech enhancement as the pre-processor of automatic speech recognition system. The improvements are based on using optimally modified log-spectra (OM-LSA) gain function with a non-causal a priori signal-to-noise ratio (SNR) estimation. Experimental results show that the proposed method gives better perceptual evaluation of speech quality score (PESQ) and lower log-spectral distance, and also better word accuracy. In the enhancement system, parameters was turned for automatic speech recognition.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2003.06a
• /
• pp.46-49
• /
• 2003

The use of confidence measures for word/utterance verification has become art essential component of any speech input application. Confidence measures have applications to a number of problems such as rejection of incorrect hypotheses, speaker adaptation, or adaptive modification of the hypothesis score during search in continuous speech recognition. In this paper, we present a new utterance verification method using vowel string. Using subword HMMs of VCCV unit, we create anti-models which include vowel string in hypothesis words. The experiment results show that the utterance verification rate of the proposed method is about 79.5%.

• MALSORI
• /
• no.60
• /
• pp.165-180
• /
• 2006

In this paper, we present several confidence measures (CM) for speech recognition systems to evaluate the reliability of recognition results. We propose heuristic CMs such as mean log-likelihood score, N-best word log-likelihood ratio, likelihood sequence fluctuation and likelihood ratio testing(LRT)-based CMs using several types of anti-models. Furthermore, we propose new algorithms to add weighting terms on phone-level log-likelihood ratio to merge word-level log-likelihood ratios. These weighting terms are computed from the distance between acoustic models and knowledge-based phoneme classifications. LRT-based CMs show better performance than heuristic CMs excessively, and LRT-based CMs using phonetic information show that the relative reduction in equal error rate ranges between $8{\sim}13%$ compared to the baseline LRT-based CMs. We use the support vector machine to fuse several CMs and improve the performance of utterance verification. From our experiments, we know that selection of CMs with low correlation is more effective than CMs with high correlation.

• Journal of Oriental Neuropsychiatry
• /
• v.24 no.1
• /
• pp.27-44
• /
• 2013

Objectives : This study was designed to investigate the effects of Gastrodia elata Blume on the improvement of memory. Methods : This study was a 12 week, double blind, comparative clinical study. There were eligible who worked with a group of healthy seniors, all 60 years of age or older. 50 subjects were randomized either to Gastrodia elata Blume in powder form and steep in hot water or placebo. We measured the faculty of memory by using K-DRS, MMSE-K, Digit Span, Letter Fluency Test, Word List Memory Test, and the Trail Making Test, and after 12 weeks we measured the faculty of memory again using the same methods. Results : Gastrodia elata Blume steeps in the hot water group significantly increased. Initiation, perseveration level, and Memory level of K-DRS and MMSE-K score. There were no considerable differences between three groups in Digit Span and Trail Making Test score. Gastrodia elata Blume group showed significant advances in Letter Fluency Test and recognition of Word List Memory Test. Conclusions : The results suggest that Gastrodia elata Blume may have positive effects on memory improvement and function of the frontal lobe activation.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.3
• /
• pp.471-476
• /
• 2016

This paper presents our study on speech recognition performance prediction. Our initial study shows that a combination of speech quality measures effectively improves correlation with Word Error Rate (WER) compared to each speech measure alone. In this paper we demonstrate a new combination of various types of speech quality measures shows more significantly improves correlation with WER compared to the speech measure combination of our initial study. In our study, SNR, PESQ, acoustic model score, and MFCC distance are used as the speech quality measures. This paper also presents our speech database verification system for speech recognition employing the speech measures. We develop a WER prediction system using Gaussian mixture model and the speech quality measures as a feature vector. The experimental results show the proposed system is highly effective at predicting WER in a low SNR condition of speech babble and car noise environments.

• Dementia and Neurocognitive Disorders
• /
• v.23 no.3
• /
• pp.136-145
• /
• 2024

Background and Purpose: We developed a new digital cognitive assessment called Seoul Cognitive Status Test (SCST), formerly called Inbrain Cognitive Screening Test. The purpose of this study was to validate the clinical utility of the SCST by comparing its scores of those with subjective cognitive decline (SCD), amnestic mild cognitive impairment (aMCI), and dementia diagnosed by the Korean version of the Consortium to Establish a Registry for Alzheimer's Disease Assessment Packet (CERAD-K). Methods: All participants (n=296) who completed the CERAD-K, SCST, and Instrumental Activities of Daily Living tests were included in this study. Total score, cognitive domain scores, and subtest scores of the SCST were compared among the 3 groups (SCD, aMCI, and dementia). Additionally, correlations between SCST and CERAD-K subtests were examined. Results: Cognitive domain scores and total score of the SCST showed significant differences among the three groups, with scores being the highest in the order of SCD, aMCI, and dementia (p<0.001). Most subtests of the SCST also showed higher scores in the order of SCD, aMCI, and dementia (p<0.001). However, SCD and aMCI groups showed no significant differences in scores of the Phonemic Word Fluency Test (p=0.083) or Korean Trail Making Test-Elderly version Part A (p=0.434). Additionally, there was no significant difference in the score of Place Recognition (p=0.274) of the Word-Place Association Test between aMCI and dementia groups. Conclusions: In conclusion, differences in total score, cognitive domain scores, and subtest scores of the SCST among the 3 groups of participants diagnosed using CERAD-K confirm the clinical utility of the SCST for cognitive assessment.

• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.327-329
• /
• 2017

개체명 인식(Named Entity Recognition) 시스템은 문서에서 인명(PS), 지명(LC), 단체명(OG)과 같은 개체명을 가지는 단어나 어구를 해당 개체명으로 인식하는 시스템이다. 개체명 인식 시스템을 개발하기 위해 딥러닝 기반의 워드 임베딩(word embedding) 자질과 문장의 형태적 특징 및 기구축 사전(lexicon) 기반의 자질 구성 방법을 제안하고, bi-directional LSTM, CNN, CRF과 같은 모델을 이용하여 구성된 자질을 학습하는 방법을 제안한다. 실험 데이터는 2017 국어 정보시스템 경진대회에서 제공한 2016klpNER 데이터를 이용하였다. 실험은 전체 4258 문장 중 학습 데이터 3406 문장, 검증 데이터 426 문장, 테스트 데이터 426 문장으로 데이터를 나누어 실험을 진행하였다. 실험 결과 본 연구에서 제안하는 모델은 BIO 태깅 방식의 개체 청크 단위 성능 평가 결과 98.9%의 테스트 정확도(test accuracy)와 89.4%의 f1-score를 나타냈다.

• 한국어정보학회:학술대회논문집
• /
• 2017.10a
• /
• pp.327-329
• /
• 2017

개체명 인식(Named Entity Recognition) 시스템은 문서에서 인명(PS), 지명(LC), 단체명(OG)과 같은 개체명을 가지는 단어나 어구를 해당 개체명으로 인식하는 시스템이다. 개체명 인식 시스템을 개발하기 위해 딥러닝 기반의 워드 임베딩(word embedding) 자질과 문장의 형태적 특징 및 기구축 사전(lexicon) 기반의 자질 구성 방법을 제안하고, bi-directional LSTM, CNN, CRF과 같은 모델을 이용하여 구성된 자질을 학습하는 방법을 제안한다. 실험 데이터는 2017 국어 정보시스템 경진대회에서 제공한 2016klpNER 데이터를 이용하였다. 실험은 전체 4258 문장 중 학습 데이터 3406 문장, 검증 데이터 426 문장, 테스트 데이터 426 문장으로 데이터를 나누어 실험을 진행하였다. 실험 결과 본 연구에서 제안하는 모델은 BIO 태깅 방식의 개체 청크 단위 성능 평가 결과 98.9%의 테스트 정확도(test accuracy)와 89.4%의 f1-score를 나타냈다.

• Speech Sciences
• /
• v.15 no.3
• /
• pp.79-94
• /
• 2008

The purpose of this study was to compare phonological awareness ability of students with Down Syndrome(DS) and typically developing children(TD). TD and DS were equal the reading abilities(reading recognition). The subject were 10 DS and 10 TD, and were examined by test of phonological awareness. The test of phonological awareness was composed according to phonological units(word, syllable, phoneme) and task types(deletion, discrimination, blending). The results obtained in this study were as follows: The total score of phonological awareness ability of DS were significantly lower than TD. And the score of phonological awareness ability according to phonological units and task types were significantly lower than TD. But both DS and TD performed better on phonological deletion and blending task than discrimination. TD and DS represented different correlation between task types and phonological units. This means that TD performed better on all types of tasks and phonological units than DS.

• Journal of KIISE
• /
• v.44 no.8
• /
• pp.813-821
• /
• 2017

Entity Linking find the meaning of an entity mention, which indicate the entity using different expressions, in a user's query by linking the entity mention and the entity in the knowledge base. This task has four challenges, including the difficult knowledge base construction problem, multiple presentation of the entity mention, ambiguity of entity linking, and NIL entity recognition. In this paper, we first construct the entity name dictionary based on Wikipedia to build a knowledge base and solve the multiple presentation problem. We then propose various methods for NIL entity recognition and solve the ambiguity of entity linking by training the support vector machine based on several features, including the similarity of the context, semantic relevance, clue word score, named entity type similarity of the mansion, entity name matching score, and object popularity score. We sequentially use the proposed two methods based on the constructed knowledge base, to obtain the good performance in the entity linking. In the result of the experiment, our system achieved 83.66% and 90.81% F1 score, which is the performance of the NIL entity recognition to solve the ambiguity of the entity linking.