摘要
为获得图像中光滑边界的稀疏表达和克服传统量化离散小波系数造成的边界失真,提出了一种基于边界的提升变换。其主要思想是将图像分为边界带和剩余区域两部分,然后对这两部分分别应用一维和二维基于边界的小波变换。边缘的前处理是这种方法的一个重要部分。为了提高编码性能,对两条平行或近似平行的边界线用一条脊线表示,并且对脊线和边界线按照光滑尖锐比进行修剪。最后根据边界小波变换系数的特点给出了合适的编码方法。本方法针对富含边界的图像在峰值信噪比和视觉质量方面比JPEG2000均有明显的提高。
An edge-based lifting scheme was proposed to give sparse representation for smooth edges in images and also to overcome artifacts around edges caused by quantizing the conventional discrete wavelet coefficients.The main idea is to divide an image into two parts：the image bands along edges and the rest of the image,then one-dimensional edge-based wavelet transform（EWT） and two-dimensional EWT are used for coding the two parts respectively.The preprocessing of the edges is an important part of our approach.To improve the coding performance,we replaced two approximately parallel and close edge segments by one ridge line.We also pruned the detected ridge lines and edges by their smoothness-to-sharpness ratio.Lastly proper coding scheme was given based on the characteristics of EWT coefficients.For images which are rich in edges,our method achieves encouraging results in both peak signal-to-noise ratio and visual quality,compared with JPEG2000.
出处
《计算机科学》
CSCD
北大核心
2012年第2期291-296,共6页
Computer Science
关键词
图像压缩
小波变换
稀疏表达
视觉质量
边界保持
Image compression
Wavelet transform
Sparse representation
Visual quality
Edge preservation
