http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
고형화 光云大學校 1989 論文集 Vol.18 No.-
In this paper, a new objective evaluation method of picture quality is proposed. Utilizing human visual property, an impairment noise, which is the sum of blocky noise in the edge and impulsive noise in the flat region, is the criterion of the evaluation. Human eye is sensitive to the jagged edge. This may be called blocky noise. But, in the flat region, eye is sensitive to the impulsive error. This may be called impulsive noise. On the basis of the two noises, a new impairment noise is generated. The algorithm needs two thresholds, T�and T�. This is the factor to be given by the evaluator. An optimum T�is between 0.2 and 0.4 and an optimum T�is between 2.5 and 4.5 in the case of the simple image. However, in the case of the complex image, T�is between 0.5 and 0.7, and T�is between 1.0 and 2.0 This is obtained by numerous simulation results.
Hartley 변환과 신경회로망을 이용한 필기체 한글인식에 관한 연구
고형화 光云大學校 1992 論文集 Vol.21 No.-
In this paper, the recognition of handwritten Korean characters is described. At first, input character is converted into binary pattern. Boundary pattern is obtained by extracting edges on the binary pattern, and then Hangul syllabics are separated from character pattern. In order to use the curved feature of Hangul syllabics, depth informations scanned from outside to boundary of pattern are transformed by discrete Hartley transform. For the purpose of complementing the feature above, descriptors obtained from boundary coordinates are used. The features above are used as input vectors of multilayer-perceptron using modified learning algorithm. In the experiment, twenty-four syllabics and twenty-five characters are obtained from forty persons, respectively. From the result of this experiment, twenty-four syllabics and twenty-five characters are obtained from forty persons, respectively. From the result of this experiment, invariance about the position and size of syllabics and the width of stroke is verified.
영상압축을 위한 프레임간 Template Matching Block Truncation Coding에 관한 연구
高亨和 光云大學校 1986 論文集 Vol.15 No.-
디지털 영상신호 의 데이터壓縮을 위한 새로운 template matching block truncation coding(TMBTC)을 提案하였다. 旣存의 BTC에 北해 4∼7배의 데이터압축률을 이루었고 화질면에서 큰 劣化가 없음을 확인하였다. 데이터 전송율을 BTC가 2.0bits/pel인데 비해, TMBTC의 데이터전송율은 0.3bit/pel이 된다. 컴퓨터 시뮬레이션을 통해 그의 우수성을 확인하였다. To compress digital picture data, a new template matching block truncation coding(TMBTC) method is proposed. Achieved data rate in TMBTC is 1/4∼1/7 of the data rate in BTC. However, the picture quality is hardly degraded. The compressed data rate of TMBTC is 0.3bit/pel, however, BTC has 2.0bit/pel. The excellency of this algorithm is shown by computer simulation.
고형화 ( Hyung Hwa Ko ),서석용 ( Seok Yong Seo ) 한국항행학회 2015 韓國航行學會論文誌 Vol.19 No.6
2진(binary) MQ 산술부호화는 최근 들어 멀티미디어 압축 표준시스템에 기본 엔트로피 방식으로 사용되고 있다. MQ 산술부호기는 JBIG2와 JPEG2000에 적용되면서 압축성능을 인정받기 시작했다. 최근에 차세대 동영상 부호화 표준인 HEVC (highefficiency video coding)에는 산술부호화가 단일 엔트로피부호화로 채택되면서 그 중요성이 커지고 있다. 기존의 2진 MQ 산술부호기는 RANGE(구간)을 분할하는 과정에서 곱셈을 없애면서 근사화 방법을 사용하고 있다. 이 경우 MPS/LPS의 구간이 뒤바뀌는 경우가 발생하며 출력비트가 늘어날 수 있다. 본 논문에서는 이러한 문제점을 완화하기 위하여 근사식을 사용하는 대신에 룩업테이블 형태로 AQe의 값을 양자화하여 계산에 적용하는 방법을 제안하고자 한다. 제안한 방법의 압축 성능을 실험을 통해 확인한결과, 2진영상 압축표준 방식인 JBIG2의 경우 약 4%의 압축율의 개선을 보였다. 정지영상 압축표준인 JPEG2000의 경우 약 1%정도의 개선을 가져왔다. 룩업테이블을 사용하기 때문에 계산량이 기존방법에 비해 늘지 않는다. Binary MQ arithmetic coding is widely used recently as a basic entropy coder in multimedia coding system. MQ coder esteems high in compression efficiency to be used in JBIG2 and JPEG2000. The importance of arithmetic coding is increasing after it is adopted as an unique entropy coder in HEVC standard. In the binary MQ coder, arithmetic approximation without multiplication is used in the process of recursive subdivision of range interval. Because of the MPS/LPS exchange activity happened in MQ coder, output byte tends to increase. This paper proposes an enhanced binary MQ arithmetic coder to make use of a lookup table for AQe using quantization skill in order to reduce the deficiency. Experimental results show that about 4% improvement of compression in case of JBIG2 for bi-level image compression standard. And also, about 1% improvement of compression ratio is obtained in case of lossless JPEG2000 coding. For the lossy JPEG2000 coding, about 1% improvement of PSNR at the same compression ratio. Additionally, computational complexity is not increasing.