• 말소리와 음성과학
• /
• 제11권2호
• /
• pp.57-64
• /
• 2019

본 논문에서는 음성 합성을 위한 오픈소스 시스템인 Merlin 툴킷을 이용하여 한국어 TTS 시스템을 구성한다. TTS 시스템에서 HMM 기반의 통계적 음성 합성 방식이 널리 사용되고 있는데, 이 방식에서 문맥 요인을 포함시키는 음향 모델링 구성의 한계로 합성 음성의 품질이 저하된다고 알려져 있다. 본 논문에서는 여러 분야에서 우수한 성능을 보여 주는 심층 신경망 기법을 적용하는 음향 모델링 아키텍처를 제안한다. 이 구조에는 전연결 심층 피드포워드 신경망, 순환 신경망, 게이트 순환 신경망, 단방향 장단기 기억 신경망, 양방향 장단기 기억 신경망 등이 포함되어 있다. 실험 결과, 문맥을 고려하는 시퀀스 모델을 아키텍처에 포함하는 것이 성능 개선에 유리하다는 것을 알 수 있고, 장단기 기억 신경망을 적용한 아키텍처가 가장 좋은 성능을 보여주었다. 그리고 음향 특징 파라미터에 델타와 델타-델타 성분을 포함하는 것이 성능 개선에 유리하다는 결과가 도출되었다.

• 대한음성학회지:말소리
• /
• 제48호
• /
• pp.57-67
• /
• 2003

This paper presents a systematic description of exceptional pronunciations for automatic Korean pronunciation generation. An automatic pronunciation generator in Korean is an essential part of a Korean speech recognition system and a TTS (Text-To-Speech) system. It is composed of a set of regular rules and an exceptional pronunciation dictionary. The exceptional pronunciation dictionary is created by extracting the words that have exceptional pronunciations, based on the characteristics of the words of exceptional pronunciation through phonological research and the systematic analysis of the entries of Korean dictionaries. Thus, the method contributes to improve performance of automatic pronunciation generator in Korean as well as the performance of speech recognition system and TTS system in Korean.

• 대한음성학회지:말소리
• /
• 제46호
• /
• pp.51-64
• /
• 2003

In order to produce high quality synthesized speech, it is very important to get an accurate grapheme-to-phoneme conversion and prosody model from texts using natural language processing. Robust preprocessing for non-Korean characters should also be required. In this paper, we analyzed Korean texts using a morphological analyzer, part-of-speech tagger and syntactic chunker. We present a new grapheme-to-phoneme conversion method for Korean using a hybrid method with a phonetic pattern dictionary and CCV (consonant vowel) LTS (letter to sound) rules, for unlimited vocabulary Korean TTS. We constructed a prosody model using a probabilistic method and decision tree-based method. The probabilistic method atone usually suffers from performance degradation due to inherent data sparseness problems. So we adopted tree-based error correction to overcome these training data limitations.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2003년도 5월 학술대회지
• /
• pp.87-90
• /
• 2003

In order to produce high quality synthesized speech, it is very important to get an accurate grapheme-to-phoneme conversion and prosody model from texts using natural language processing. Robust preprocessing for non-Korean characters should also be required. In this paper, we analyzed Korean texts using a morphological analyzer, part-of-speech tagger and syntactic chunker. We present a new grapheme-to-phoneme conversion method, i.e. a dictionary-based and rule-based hybrid method, for unlimited vocabulary Korean TTS. We constructed a prosody model using a probabilistic method and decision tree-based method.

• Journal of Pharmaceutical Investigation
• /
• 제33권2호
• /
• pp.141-144
• /
• 2003

The transdermal therapeutic system (TTS) of scopolamine has various advantages over its oral dosage forms. The ideal scopolamine TTS requires high skin permeation rate in short time after it is applied on the skin. In order to increase the initial skin permeation rate of scopolamine from TTS, various permeation enhancers were employed. Enhancers employed were fatty acids (oleic and linolenic acids), cyclic monoterpenes (menthol, camphor, cineole and limonene) and others (isopropyl myristate, sodium lauryl sulfate and glyceryl monostearate). The concentration of enhancers in the base were fixed to 5% (w/w). While fatty acids had little enhancing effect on the skin permeation of scopolamine, cyclic monoterpenes, isopropyl myristate and sodium lauryl sulfate resulted in $1.5{\sim}2.6-fold$ higher skin permeation rate of the drug compared to the control. However, lag time was not affected by enhancers studied.

• BMB Reports
• /
• 제48권5호
• /
• pp.283-288
• /
• 2015

We found that resveratrol enhances interferon (IFN)-γ-induced tryptophanyl-tRNA-synthetase (TTS) expression in bone marrow-derived dendritic cells (BMDCs). Resveratrol-induced TTS expression is associated with glycogen synthase kinase-3β (GSK-3β) activity. In addition, we found that resveratrol regulates naive CD8⁺ T-cell polarization by modulating GSK-3β activity in IFN-γ-stimulated BMDCs, and that resveratol induces upregulation of TTS in CD8⁺ T-cells in the in vivo tumor environment. Taken together, resveratrol upregulates IFN-γ-induced TTS expression in a GSK-3β-dependent manner, and this TTS modulation is crucial for DC-mediated T-cell modulation. [BMB Reports 2015; 48(5): 283-288]

• 한국컴퓨터정보학회논문지
• /
• 제26권12호
• /
• pp.105-110
• /
• 2021

본 논문에서는 STT(Speech-to-Text), TTS(Text-to-Speech) API와 가속도 센서 기반의 청각 장애인의 의사소통을 도와주는 안드로이드 모바일 애플리케이션을 설계하고 구현한다. 이 애플리케이션은 청각 장애인의 대화 상대가 말하는 것을 마이크로 녹음하고 STT API를 이용하여 텍스트로 변환하여 청각 장애인에게 보여주는 기능을 제공한다. 또한, TTS API를 이용하여 청각 장애인이 문자를 입력하면 음성으로 변환하여 대화 상대에게 들려준다. 청각 장애인이 스마트폰을 흔들면 이 애플리케이션이 실행하도록 가속도 센서 기반의 백그라운드 서비스 기능을 제공한다. 본 논문에서 구현한 애플리케이션은 청각 장애인들이 다른 사람과 의사소통을 할 때 영상통화로 수화를 이용하지 않고 쉽게 대화할 수 있는 기능을 제공한다.

• 말소리와 음성과학
• /
• 제13권4호
• /
• pp.47-53
• /
• 2021

Autoregressive한 TTS 모델은 불안정성과 속도 저하라는 본질적인 문제를 안고 있다. 모델이 time step t의 데이터를 잘못 예측했을 때, 그 뒤의 데이터도 모두 잘못 예측하는 것이 불안정성 문제이다. 음성 출력 속도 저하 문제는 모델이 time step t의 데이터를 예측하려면 time step 1부터 t-1까지의 예측이 선행해야 한다는 조건에서 발생한다. 본 연구는 autoregression이 야기하는 문제의 대안으로 end-to-end non-autoregressive 가속 TTS 모델을 제안한다. 본 연구의 모델은 Tacotron 2 - WaveNet 모델과 근사한 MOS, 더 높은 안정성 및 출력 속도를 보였다. 본 연구는 제안한 모델을 토대로 non-autoregressive한 TTS 모델 개선에 시사점을 제공하고자 한다.

• Journal of Trauma and Injury
• /
• 제35권4호
• /
• pp.248-254
• /
• 2022

Purpose: To analyze the data of trauma patients with undetected injuries at the time of initial resuscitation during the primary and secondary surveys. Methods: We retrospectively reviewed the medical records of 807 patients who were hospitalized at the National Trauma Center, Seoul, Korea from June 1, 2019 to June 30, 2021. Results: In trauma patients with an Injury Severity Score ≥16 accounted for 27.5% in the non-missed injury group (non-MIG), but this rate was considerably higher at 71.2% in MIG. The mean hospitalization longer in MIG (50.90±39.56) than in non-MIG (24.74±26.11). The proportion of patients with missed injuries detected through tertiary trauma survey (TTS) was 28 patients (23.5%) within 24 hours, 90 patients (75.6%) after 24 hours to before discharge. The majority of missed injuries were fractures (82.4%) and ligament tears (8.4%), which required consultation with the orthopedic department. The final diagnoses of missed injuries were confirmed by computed tomography (44.5%), magnetic resonance imaging (19.3%), X-ray (19.3%), bone scan (11.8%), and physical examination (5.0%). Conclusions: TTS is considered a useful process for detecting missed injuries that were not identified at the time of initial resuscitation in the primary and secondary surveys. In the future, to detect missed injuries quickly, it is necessary to develop a suitable TTS program for each trauma center. In addition, further research is needed to verify the effectiveness of the protocolized TTS and survey chart to improve the effectiveness of TTS.

• 시스템엔지니어링학술지
• /
• 제6권2호
• /
• pp.29-35
• /
• 2010

In general, an Architecture of a system is embodied as applied results of a requirement analysis of a system in early development phase. These efforts play a important role in analyzing and understanding a system considering operational, functional, and physical view and deriving a correct solution before developing the system. In this paper, the architecture of the Tube Transportation System(TTS) known as the new transportation system in Railway Domain is depicted by performance parameter has already developed. The existing performance parameters are shown by a variety of types with many meanings rather than types of general requirements refined. As these early performance parameters have analyzed and complemented to a level of requirement by requirement managers and other domain specialists, the architecture of the Tube Transportation System was developed systematically and then system requirements will be drawn up definitely. The presented architecture will provide a framework of developing a TTS and also offer an information in performance analysis of TTS.