基于贝叶斯阈值的多小波基联合去噪
Jointed Multi-Wavelet Bases Based Bayesian Thresholding for Image Denoising
-
摘要: 本文用贝叶斯阈值法来确定阈值,提出在不同子带和不同方向上使用不同的阈值,结合改进的软阈值函数,并采用2个小波基加权联合去噪.实验证明,此法与Visushrink、Sureshrink、Bayesshrink及Wiener滤波相比,在去噪的同时保留了更多的图像纹理,峰值信噪比也提高0.1~1.45dB不等,充分说明了该方法的有效性.Abstract: This paper proposed different thresholds based on Bayesian framework for different subbands and orientations combined with improved soft threshold function.Meanwhile, this paper used two wavelets to get weighted average final image.Experiments show that this method remains more grain and raise PSNR form 0.1 to 1.45dB.It was effective when comparing with Visushrink, Sureshrink, Bayesshrink and Wiener filtering.
下载:
