• Journal of Biomedical Engineering Research
• /
• v.39 no.6
• /
• pp.278-283
• /
• 2018

Nowdays, many people suffer from the neck pain due to forward head posture(FHP) and text neck(TN). To assess the severity of the FHP and TN the craniovertebral angle(CVA) is used in clinincs. However, it is difficult to monitor the neck posture using the CVA in daily life. We propose a new method using the cervical flexion angle(CFA) obtained from a wearable sensor to monitor neck posture in daily life. 15 participants were requested to pose FHP and TN. The CFA from the wearable sensor was compared with the CVA observed from a 3D motion camera system to analyze their correlation. The determination coefficients between CFA and CVA were 0.80 in TN and 0.57 in FHP, and 0.69 in TN and FHP. From the monitoring the neck posture while using laptop computer for 20 minutes, this wearable sensor can estimate the CVA with the mean squared error of 2.1 degree.

• MALSORI
• /
• no.46
• /
• 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.

• Phonetics and Speech Sciences
• /
• v.4 no.2
• /
• pp.51-59
• /
• 2012

The purpose of this study is to examine the pitch pattern of Korean accentual phrases produced by Chinese Korean-learners in the reading of a Korean text. Korean accentual phrase is determined by a specific F0 contour. And the pitch contour of APs differ depending on their length and the nature of initial segment. In order to examine if Chinese speakers are also aware such a phonetic properties, we have examined the AP pitch contours produced by 15 Chinese speakers differing in proficiency, and compared them to pitch contours produced by six Korean native speakers. The results show that Chinese speakers' pitch errors were observed in initial segment-tone interaction and in type of pitch patterns. However, even though Chines speakers produced the same type of pitch patterns, internal tonal modulation differs from native speakers. Finally, on the basis of theses results, we proposed a teaching method that visualizes the F0 contour.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.608-612
• /
• 2022

딥러닝의 발전 이후, 다양한 분야에서는 딥러닝을 이용해 이전에 어려웠던 작업들을 해결하여 사용자에게 편의성을 제공하고 있다. 하지만 아직 딥러닝을 통해 이상적인 서비스를 제공하는 데는 어려움이 있다. 특히, 음성 인식 작업에서 음성 양식에서 이용 방안에 대하여 다양성을 제공해주는 음성을 텍스트로 전환하는 Speech-To-Text(STT)은 문장 결과가 이상치에 달하지 못해 오류가 나타나게 된다. 본 논문에서는 STT 결과 보정을 문법 교정으로 치환하여 종단에서 올바른 토큰들을 조합하여 성능 향상을 하기 위해 각 토큰 별 품질 평가를 진행하는 모델을 한국어에서 적용하고 성능의 향상을 확인한다.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.5
• /
• pp.1273-1280
• /
• 1996

In order perform high accuracy recognition of text recognition systems, the recognized text must be processed through a post-processing stage using contextual information. We present a system that combines multiple knowledge sources to post-process the output of an optical character recognition(OCR) system. The multiple knowledge sources include characteristics of word, wrongly recognized types of Hangul characters, and Hangul word learning In this paper, the wrongly recognized characters which are made by OCR systems are collected and analyzed. We imput a Korean dictionary with approximately 15 0,000 words, and Korean language texts of Korean elementary/middle/high school. We found that only 10.7% words in Korean language texts of Korean elementary/middle /high school were used in a Korean dictionary. And we classified error types of Korean character recognition with OCR systems. For Hangul word learning, we utilized indexes of texts. With these multiple knowledge sources, we could predict a proper word in large candidate words.

• Journal of Biomedical Engineering Research
• /
• v.37 no.3
• /
• pp.112-118
• /
• 2016

The neck pain is fairly common occurance. Forward head posture and text neck are poor postures which may be related with neck pain but the evidence is not enough. We developed the wearable sensor which can assess the head & neck posture in daily life. Microprocessor, Bluetooth low energy, and 3-axis accelerometer, rechargeable battery and vibratior for reminding are used to implement the wearable sensor. Real-time algorithm to parameterize the posture for one epoch is implemented which classifies the posture in the epoch into three classed; dynamic, static_good posture, and static_poor posture. Also the algorithm makes reminding to its wearer to give them the prolonged poor posture is detected. The mean error of measurement was 1.2 degree. The correlation coefficient between neck angle and craniovertebral angle was 0.9 or higher in all cases. With the pilot study on text neck syndrome was also quatified. Average of neck angle were 74.3 degree during the listening in the classroom and 57.8 degree during the smartphoning. Using the wearable sensor suggested, the poor postures of forward head posture and neck neck can be detected in real-time which can remind the wearer according to his/her setting.

• The Journal of Korean Medical History
• /
• v.31 no.2
• /
• pp.9-16
• /
• 2018

Eoyakwonbang is a collection of prescriptions of Yuyaoyuan, an imperial medical bureau of China. While the first edition of this book does not exist at present, two versions printed in Korea, and one printed in Japan, which is the well-known Gyesasingan Eoyakwonbang have been passed down. Eoyakwonbang is a meaningful text for the history of medical communication between China, Japan and Korea, but research on the different versions and contents of Eoyakwonbang have been insufficient. Out of the 2 versions published in Korea, Eulheja Eoyakwonbang is different from Gapjinja Eoyakwonbang [another version in Korea] and Gwanjeong Eoyakwonbang of Japan, in that the prescriptions are organized, the ingredients are organized according to the amount of each ingredient for each formula, and the amounts are recorded in an accurate manner. On the other hand, the Gapjinja version has many mistakes in the characters and content. The Gwanjeong version has lesser mistakes in characters, but repeats the content error of the Gapjinja version. Eulheja was printed after correction based on the original version or unknown version from China. Gapjinja was re-printed based on the Gyesasingan version, while Gwanjeong seems to have used the Gapjinja version as its original script, as the mistakes made in the Gapjinja version are repeated in the Gwanjeong version as well.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.3
• /
• pp.274-284
• /
• 2002

Word segmentation errors occurring in text preprocessing often insert incorrect words into recognition vocabulary and cause poor language models for Korean large vocabulary continuous speech recognition. We propose an automatic word segmentation algorithm using Markov chains and syllable-based n-gram language models in order to correct word segmentation error in teat corpora. We assume that a sentence is generated from a Markov chain. Spaces and non-space characters are generated on self-transitions and other transitions of the Markov chain, respectively Then word segmentation of the sentence is obtained by finding the maximum likelihood path using syllable n-gram scores. In experimental results, the algorithm showed 91.58% word accuracy and 96.69% syllable accuracy for word segmentation of 254 sentence newspaper columns without any spaces. The algorithm improved the word accuracy from 91.00% to 96.27% for word segmentation correction at line breaks and yielded the decomposition accuracy of 96.22% for compound-noun decomposition.