http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
김선형,이순흠,김기석,안덕상,Kim, Sun-Hyung,Lee, Sun-Heum,Kim, Gee-Seog,Ahn, Deog-Sang 한국정보처리학회 2002 정보처리학회논문지D Vol.9 No.5
This paper discusses techniques for embedding data into 3D polygonal models of geometry. Much researches of Watermarking had been gone as element technology of DRM (digital rights management). But, few research had gone to 3D polygonal model. Most research is limited at text document, 2D image, animation, music etc. RP system is suitable a few production in various goods species, and it is used much in industry to possible reason that produce prototype and find error or incongruent factor at early stage on design in product development childhood. This paper is research about method that insert watermark in STL ( stereolithography) file that have 3D shape model. Proposed algorithm inserts watermark in normal vector region and facet's interior region of 3D shape data. For this reason, 3D shape does not produce some flexure and fulfill invisibility of watermark. Experiment results that insert and extract watermark in normal netter region and facet's Interior region of 3D shape data by proposed algorithm do not influence entirely in 3D shape and show that insertion and extraction of watermark are possible. 워터마킹은 DRM(digital rights management)의 요소 기술로서 많은 연구가 진행되어 왔다. 하지만, 3D 신속 조형 기술(RP : rapid proto-typing)에 적용된 연구는 찾아보기 어렵다. 대부분의 연구가 텍스트 문서, 2D 이미지, 동영상, 음악 등의 컨텐츠에 한정되어 있다. RP 시스템은 다품종 소량 생산에 적합하고, 시제품을 제작하여 제품개발 초기 단계에서 설계상의 오류나 부적합한 요인을 조기에 발견하는 것이 가능한 까닭에 산업 현장에 많이 사용되고 있다. 본 논문은 3D 형상 모델을 가지고 있는 STL(stereolithography) 파일에 워터마크를 삽입하는 방법에 관한 연구이다. 제안된 알고리즘은 3D 형상 데이터의 법선 영역과 패싯 뒷면 영역에 워터마크를 삽입한다. 이런 까닭에 3D 형상에는 어떠한 왜곡도 발생하지 않으며 워터마크의 비가시성(invisibility)을 충족한다. 제안된 알고리즘으로 3D 형상 데이터의 법선 영역과 패싯 뒷면 영역에 워터마크 데이터의 삽입과 추출이 가능함을 보여주었으며, 본 연구의 실험 결과는 fragile 워터마킹과 robust 워터마킹과의 이용가능성을 제시하였다.
Off-Line 필기체 문자 인식을 위한 효율적인 전처리기의 구성
김기석,홍인식,천인국 순천향대학교 부설 산업기술연구소 2000 순천향 산업기술연구소논문집 Vol.6 No.1
The purpose of this paper is to increase the cognition rates on hand-written characters. Unlike printed characters, the cognition of hand-written characters requires an algorithm which takes into consideration of the individual differences. Characters are basically a set of points. They can be represented in terms of feature parameters such as glyph vector that means sum of points per specific area in input character image. Then all characters can be represented by glyph multiplied by the feature parameters respectively. In this paper we propose a method for character cognition of off-line hand-written in terms of BP algorithm inference. That method makes improving the cognition quality of hand-written characters in speed and cognition rates. Usually improving the cognition quality of hand-written characters requires many preprocessing steps. These many steps tend to make the overall recognition system slower and less efficient. Therefore, an effective preprocessing, midprocessing and postprocessing scheme is proposed.
김기석,안덕상,김선형,천인국 순천향대학교 부설 산업기술연구소 2002 순천향 산업기술연구소논문집 Vol.8 No.1
This paper discusses techniques of embedding data into 3D polygonal models of geometry. Many researches of Watermarking have been done as element technology of DRM(digital rights management) and are limited to text document, 2D image, animation, music etc. But, few researches have been done on 3D polygonal model. this paper proposes a method that insers watermark in STL(standard transfer language)file. The proposed algorithm insers multi-bit watermark information into the surface normal vector of STL file without distorting the original 3D geometric information. And also it does not produce some flexure of 3D shape and fulfills the invisibility of watermark. Experiment results, that insert and extract watermark in normal vector region of 3D polygonal data by proposed algorithm, do not distort the original 3D shape and show that successful insertion and extraction operation of watermark are possible.
신경 회로망과 진화 프로그래밍을 이용한 Off-Line 필기체 문자 인식
김기석,홍인식,천인국 순천향대학교 부설 산업기술연구소 2000 순천향 산업기술연구소논문집 Vol.6 No.1
Generally speaking, pattern cognition system would be run since neural network is initialized by trained data. Vector, that is obtained from preprocessor, is input of neural network, and result is out of neural network. In case of time cost for pattern recognition, it is needed much more cost for training than cognition. The purpose of this paper is to reduce a cost for training and to increase the cognition speed. In pattern recognition, it is important to choose a recognizor as much as preprocessor. Therefore, back propagation neural network, that has a highly efficient for nonlinear problem, is used in implemented recognizor. Most neural network has a fixed training rate for training. But if it is trained with fixed training rate, oscillation and local maxima problem occurs in neural network, in the result, time cost will increase. Therefore, to improve these difficulties, we use a DTS(Dynamic Training Rate Selection) that selects training rate dynamically in run time. To realize DTS, We use EP algorithm. EP algorithm evolves a training-rate from 1st generation into 5th. Then, we can get a proper training-rate for training of neural network. On results of test, implemented system decreases time cost for training as 5~20%.
변호준,이효진,김기석,천인국 순천향대학교 부설 산업기술연구소 2001 순천향 산업기술연구소논문집 Vol.7 No.2
To prevent from copying digital contents and to protect a copyright, it have used a watermark system which is a component of DRM(Digital Rights Management), When a watermark is extracted, a conventional watermark method needs original image. This needs additional cost, and it can't extract a watermark without original image.In this work, we propose a method to extract a watermark without original image. The proposed method uses a wavelet transformation to embed and extract a watermark. We experiment two cases for testing the proposed method, one thing is a extracting watermark using original watermark information, the other is a extracting watermark not using original watermarking information in extracting process. The proposed method, compared to conventional method which need a original image and a watermark image, offers a lot of convenience for managing digital contents.
공용해,엄태준,방재철,김기석 순천향대학교 부설 산업기술연구소 2001 순천향 산업기술연구소논문집 Vol.7 No.1
We developed an algorithm that effectively extracts slicing points from STL 3D models for a new RP system which can fabricate various materials. Since an STL file has a simple data structure and contains much embiguity, we considered all the cases between facet positions and slicing planes for effective slice extraction from the STL file, and reduced the overall processing time. The effectiveness and efficiency of the proposed slicing method was tested by many STL 3D samples.