프랙탈 코딩을 이용한 영상 압축 기법에 관한 연구 = A Study on the Fractal Image Compression Based on the Iterated Function System

다국어 초록 (Multilingual Abstract)

Abstract
In this paper, a conceptual linkge between existing fractal image coding algorithms and classical fractal studies is made. Proposed fractal image coding is accomplished by the use of scale property, which represents the self-affinity of fractal and is defined on fractional white Gaussian noise process(FWGNP) and fractional Brownian motion process (FBMP).
For a quantitative fractal measure, an image is modeled by fractional differencing model(FDM) which is Known as a discrete version of FWGNP. Since differencing parameter of the model is related to Hurst coefficient, H, the range of this parameter makes an image categorized into a sample path of FWGNP or FBMP and the parameter value plays a central role in calculating the contractivity of each scale theorem defined on FWGNP and FBMP.
In order to prove the feasibility of our fractal image coding scheme, it is compared with existing fractal image compression algorithms proposed by Jacquin and Fisher et. al. Result shows dramatically reduced computational burden, while reconstructed image quality of our scheme is nearly the same or higher than that of the existing algorithms.

목차 (Table of Contents)

• 차례
• Ⅰ. 서론
• Ⅱ. 반복 함수계 이론
• Ⅲ. 소수차 브라운 운동
• Ⅳ. 소수차 차분 모델
• Ⅴ. 파라미터 추정 방법
• Ⅵ. 모의 실험 및 결과 고찰
• Ⅶ. 결론
• 참고문헌