멀티웨이블릿 변환영역에서 계수정규화를 이용한 Soft-Threshold 기법의 영상신호 잡음제거
Image Signal Denoising by the Soft-Threshold Technique Using Coefficient Normalization in Multiwavelet Transform Domain
0
조회
0
다운로드
0
인용
0
참고문헌
초록
웨이블릿 축소 기법으로 영상신호의 잡음을 제거할 때, 웨이블릿 계수 들이 상관관계를 갖는 경우 잡음제거 효과가 저하된다. 멀티웨이블릿 변환된 계수 들은 사전 필터의 영향으로 상관관계를 갖게 된다. 이러한 문제점을 해결하기위해 V Sterela에 의해 Universal 경계 값 적용을 위한 사전 필터를 새로 설계하거나 가중 값을 적용하는 기법이 제시되었다. 본 논문에서는 멀티웨이블릿 변환 영역에서 웨이블릿 축소 기법의 잡음제거 효과를 향상시키기 위해, 대역의 계수를 추정된 잡음편차로 나누는 계수 정규화기법을 Universal, SURE 및 GCV 경계 값에 적용하여 잡음을 제거하는 시도를 하였다. 각 경계 값들에 대한 PSNR을 비교하여 이 기법의 실용성을 확인하였다.
Abstract
In case of wavelet coefficients have correlation, in image signal denoising using wavelet shrinkage denoising method, the denoising effect for the image signal is reduced when the wavelet shrinkage denoising method is used. The coefficients of multiwavelet transform have correlation by pre-filters. To solve the degradation problem in multiwavelet transform, V Sterela suggested a new pre-filter for the Universal threshold or weighting factors to the threshold. In this paper, to improve the denoising effect in the multiwavelet transform, the coefficient normalizing method that the coefficient are divided by estimated noise deviation is adopted to the transformed multiwavelet coefficients in the course of wavelet shrinkage technique. And the thresholds of universal, SURE and GCV are estimated using normalized coefficients and tried to denoise by the wavelet shrinkage technique. We compared PSNRs of denoised images for each thresholds and confirmed the efficiency of the proposed method.
참고문헌 (20)
- "Ten Lectures on Wavelets" (1992)
- "Wavelets and Filter Banks" (1995), MA : Wellesley Cambridge Press
- "The Application of Multiwavelet Filterbanks to Image Processing", 8(4), pp.548~563
- "Fractal Function and Wavelet Expansions based on Several Functions", 78, pp.373~ (1994)
- "Short Wavelets and Matrix dilation equations", 43(1), pp.108~114
- "Ideal Spatial Adaptation by Wavelet shrinkage", 81, pp.425~ (1994)
- "Adapting to Unknown Smoothness via Wavelet Shrinkage", 90(432), pp.1200~1224
- "Wavelet Shrinkage Using Cross- validation", 58, pp.463~479 (1996)
- "Choice of the Threshold Parameter in Wavelet Function Estimation", pp.261280~ (1995), Springer-Verlug
- "Wavelet Shrinkage and Generalized Cross Validation for Image Denoising", 7(1), pp.82~92
- "Noise Reduction by Wavelet Thresholding" (2001), Springer-Verlag
- "Multiple Wavelet Threshold Estimation by Generalized Cross Validation for Images with Correlated Noiselated Noise", 8(7), pp.947~953
- "Wavelet Threshold Estimators for Data with Correlated Noise", 59, pp.319~351 (1997)
- "Multiwavelet Bases with Extra Approximation Properties", 46, pp.2898~2908
- "Design of Prefilters to Discrete Multiwavelet Transforms", 44(1), pp.25~35
- "The Discrete Multiple Wavelet Transform and Thres- holding Methods", 49, pp.2558~2561
- "Signal Image Denoising Via Wavelet Thresholding : Orthogonal and Biorthogonal Scalar and Multiple Wavelet Transforms" (1998), Imperial College, Statistics Section
- "A Study of Orthonormal Multiwavelets", 20, pp.272~298 (1996)
- "A Wavelet Tour of Signal Processing", pp.434~446 (1999), Academic Press
- "Statistical and Adaptive Signal Processing", pp.75~147 (2000), McGraw -Hill