• The Transactions of the Korea Information Processing Society
• /
• v.6 no.12
• /
• pp.3589-3596
• /
• 1999

Foreign countries have been used closed captioning for people with hearing loss. Recently some TV programs in Korea also have started to use it. Closed captioned films and video tapes of foreign countries are especially useful for people who want to learn foreign languages as an effective educational tool. In this paper, we propose a method of implementing captioning functionality to an existing streaming software. In addition we added repetitive and group selection of video segments for the new interface of our prototype system. We keep adding additional data and extending functions of our prototype system.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.04a
• /
• pp.393-396
• /
• 2010

지금까지 필기 변이를 완화하기 위한 다양한 비선형 정규화 방법들이 제안되었으며 실제 인식 시스템에서 상당한 인식률 개선 효과를 나타내었다. 그러나, 필기 한글 인식에 있어서는 필기 변이 외에도 문자간의 높은 유사도로 인해 높은 인식률을 얻는데 어려움을 겪고 있다. 한글과 같이 문자간 유사도가 높은 언어를 효과적으로 인식하기 위해서는 필기 변이를 흡수하는 것뿐 아니라, 유사 문자간의 차이를 정확히 찾아내어 그 차이점을 부각시키는 것이 요구된다. 본 논문에서는 유사 문자간의 차이점을 부각시킬 수 있는 비선형 정규화 방법을 제안한다. 기존의 비선형 정규화 방법들이 영상의 지역적 복잡도를 균일화 함으로써 정규화를 수행했던 것에 반해, 제안하는 방법에서는 유사 문자쌍의 구분에 있어 지역적 공헌도에 기반하여 영상을 정규화한다. 즉, 유사 문자쌍 구분에 공헌도가 높은 지역은 확대하고 그렇지 않은 지역은 축소한다. 그 결과, 문자간에 서로 상이한 지역을 강조 함으로써 유사 문자쌍에 대한 구분력을 높인다. 실험 결과, 제안하는 방법으로 정규화된 영상에서는 유사 문자쌍의 차이점이 확대되었으며, 문자쌍의 구분 성능 또한 향상되었다.

• Annual Conference on Human and Language Technology
• /
• 1995.10a
• /
• pp.131-136
• /
• 1995

최근들어 자연스럽게 필기된 한글을 인식함으로써 정보 입력 과정을 자동화하기 위한 오프라인 한글 글씨 인식에 관한 연구가 활발히 진행되고 있다. 오프라인 한글 글씨 인식에 관한 연구에 있어서 반드시 확보되어야 하는 연구 환경으로 대용량 오프라인 한글 글씨 데이타베이스의 구축을 들 수 있는데, 본 논문에서는 시스템공학연구소 국어공학센터의 국어 정보 베이스 개발사업의 일환으로 추진중인 오프라인 한글 글씨 데이타베이스의 구축현황에 대해 간략히 소개하고자 한다. 오프라인 한글 글씨 데이타베이스의 구축은 크게 글씨 데이타베이스 설계, 글씨 데이타 수집, 용지 스캔 및 문자 단위 분할, 데이타베이스 검증의 4 단계로 구성된다. 본 연구에서는 다양한 변형을 갖는 글씨체의 수집을 데이타베이스 구축시 가장 고려해야 할 요소로 삼았으며, 고품질의 일관성 있는 글씨 데이타베이스 구축을 위해 데이타베이스 설계 단계와 검증 단계에 많은 시간을 할애했다. 마지막으로 본 연구에서는 WWW(World Wide Web)의 HTML(Hyper Text Markup Language)을 이용하여 편리 한 사용자 인터페이스를 구현함으로써 사용자들이 쉽게 한글 글씨 영상을 검색 할 수 있음은 물론 인식 알고리즘의 개발에 사용 가능한 형태의 화일을 제공받을 수 있도록 구성하고 있다. 현재는 KS C 완성형 한글 2,350자 중에서 사용 빈도순 상위 520자에 대한 한글 글씨 1,000벌을 수집하여 명도영상 데이타베이스를 구축 중에 있으며, 향후 2년간 나머지 1,830자에 대한 한글 글씨 데이타를 수집하여 데이타베이스를 완성하고자 한다. 구축된 글씨 데이타베이스는 조만간 국내의 오프라인 한글 글씨 인식 연구자들에게 제공되어 우수한 인식 알고리즘의 개발을 위한 중요한 실험 데이타로서 사용될 예정이며, 개발된 인식 시스템에 대한 객관적인 성능 평가에 있어서도 크게 기여하여 국내의 오프라인 한글 글씨 인식에 관한 연구를 활성화시켜주는 계기가 될 것으로 기대된다.

• Proceedings of the Korea Contents Association Conference
• /
• 2006.05a
• /
• pp.267-272
• /
• 2006

The motion-graphic have been a more important factor in image illustration and typography since the development of visual culture was advanced. Therefore the visualized case of intented content with the creative typography is easily seen in television CF, movies and web pages. They suggest that variable factors such as language, time, shape, motion, color and sound should be applied and produced to motion typogaraphy. But the physiologic features as sex, age, health status, illness and physical size have an important effect in the communication process. So, the more effective result were gained than the fast-developing other mass media if these features were applied to the motion typography with semi-language speciality.

• Convergence Security Journal
• /
• v.15 no.3_1
• /
• pp.41-51
• /
• 2015

Human have more complicate and skilled ability for lying even cheat ourself. It is not easy to cheat unconscious things like sweat, eyes, or voice, but if some one cheat own self, he can cheat every of that. Lie is one of the way to spread our gene and our instinct make a lie. Every living organism even bacteria or virus use similar trick to survive. In human body, there are more complicate and profound mechanism for lying like breathe, sweat, eyes, face or voice. We can control some of that and make a fake, but it can't be perfect. Human also called 'Homo Fallax' cause we have a language and skill to lie with it. In present, we can detect lie with polygraph, but it has few weakness. So we try to use Vibraimage technology for resolve it. In this paper, we describe how to use Vibraimage for lie detection and the research history.

• Journal of the Korean earth science society
• /
• v.21 no.6
• /
• pp.655-662
• /
• 2000

According to the development of internet with Web, WBI has greatly influence on the present educational society. However, it is difficutly to design the web of the dynamic motions of graphics or animation using general programming technique based on high or low level language. Recently, Mecromedia Company supported a tool that is called Authoware which is the leading visual rich-media authoring solution for creating Web and online learning applications, to solve the problem easily. In the paper, using the the Authoware we tried to develop a web page about tidal variations due to sea level change and intertidal zone variations using the Authorware 5.1. To do this, we used the ocean survey data of Iho beach and the tidal level data based on Tidal Tables of Cheju harbor. The results showed that the Authorware was very useful to construct the simulation of tidal phenomena on web. Therefore, the Authorware can be applied to the simulation related with animation and dynamic motions for the other WBI objective.

• Archives of design research
• /
• v.15 no.3
• /
• pp.259-270
• /
• 2002

In 1995, When people read a poem, the image that a poet intends to convey to readers shows in various colors according to the status of their emotion. Poetry is a bridge as well as a text, which connects this world and the poet's world. In such relationship, the communication through Types occurs. The realm of application of modern typography is widening due to the development of the Internet and mass media, and the ways of expression of which are changing with the help of lots of softwares. So, the modern typography is re-born as an organic language which is alive, breathing. Therefore, Types has the structural character similar to that of Typography, which is a language of image, creating today's movement, time, and space. The already existing poetry contains meanings but has a descriptive structures. On contrary, compared with the former, the type appeared in Hangul Concrete Poetry., itself is a poem in another realm due to the formality native to Hangul, and which appears in non-linear structure. So, in this thesis, I will analyze the formality and non-linear structure of Hangul Typography in order to widen the realm of research on typography, which is a very meaningful trial to visualize the literature.

• Journal of Game and Entertainment
• /
• v.2 no.2
• /
• pp.61-69
• /
• 2006

Along with advance in visual culture, the importance of motion graphic has been increasingly emphasized day by day, which can maximize information delivery using image illustration and typography, graphic factors of images. In addition, we can easily see increasing cases where what a designer intends is visualized using creative typography in diverse mass media such as TV commercials, movies or web. Thanks to the effects of this trend, various ways of manufacturing works have been proposed in the field of motion typography by applying diverse factors including verbal ones, time, form, motion, colors, and sound for the purpose of expressing formless semantic notions through visual form of typography. However, physiological features such as sex, age, health status, pathological conditions, and body size can have a bigger effect on the process of real communication. Therefore, if property of quasi-verbal sound can be reflected appropriately in motion typography where communication is expressed only by visual form, it can enable people to understand what a designer intends faster and more exactly.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.4
• /
• pp.133-142
• /
• 2021

With the recent advancements of deep learning, companies such as smart home, healthcare, and intelligent transportation systems are utilizing its functionality to provide high-quality services for vehicle detection, emergency situation detection, and controlling energy consumption. To provide reliable services in such sensitive systems, deep learning models are required to have high accuracy. In order to develop a deep learning model for analyzing previously mentioned services, developers should utilize the state of the art deep learning models that have already been verified for higher accuracy. The developers can verify the accuracy of the referenced model by validating the model on the dataset. For this validation, the developer needs structural information to document and apply deep learning models, including metadata such as learning dataset, network architecture, and development environments. In this paper, we propose a description language that represents the network architecture of the deep learning model along with its metadata that are necessary to develop a deep learning model. Through the proposed description language, developers can easily verify the accuracy of the referenced deep learning model. Our experiments demonstrate the application scenario of a deep learning description document that focuses on the license plate recognition for the detection of illegally parked vehicles.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.11
• /
• pp.481-492
• /
• 2023

This study explores how to build a Korean dataset to extract information from text using generative large language models. In modern society, mixed information circulates rapidly, and effectively categorizing and extracting it is crucial to the decision-making process. However, there is still a lack of Korean datasets for training. To overcome this, this study attempts to extract information using text-based zero-shot learning using a generative large language model to build a purposeful Korean dataset. In this study, the language model is instructed to output the desired result through prompt engineering in the form of "system"-"instruction"-"source input"-"output format", and the dataset is built by utilizing the in-context learning characteristics of the language model through input sentences. We validate our approach by comparing the generated dataset with the existing benchmark dataset, and achieve 25.47% higher performance compared to the KLUE-RoBERTa-large model for the relation information extraction task. The results of this study are expected to contribute to AI research by showing the feasibility of extracting knowledge elements from Korean text. Furthermore, this methodology can be utilized for various fields and purposes, and has potential for building various Korean datasets.