摘要
基于小波变换的图像压缩算法在比特率较低时会出现重构图像边缘模糊现象.在充分考虑高频子带小波系数的分布特性对保存边缘、纹理等信息的影响之后,提出了一种新的基于势函数模糊聚类的方法来解决重构图像边缘模糊问题.实验结果显示:这种方法能较好地保存图像边缘和纹理信息,有效地减少边缘模糊现象和改善重构图像的主观质量.
The edge fuzzy phenomenon exists in the wavelet-based compressed image at low bit rates. After the statistic characteristics of each subband images’wavelet coefficients and their influence upon saving the edge and texture information of the original image are considered carefully,a new method of wavelet image compression is presented based on potential fuzzy clustering to reduce the edge fuzzy problem.The experimental results show that this method is able to keep more information of the edge and texture of the original image,considerablely reduce the edge fuzzy phenomenon and improve the subjective quality of the reconstructed image.
出处
《应用科技》
CAS
2005年第3期3-5,共3页
Applied Science and Technology
关键词
小波变换
图像压缩
模糊聚类
量化
wavelet transform
image compression
fuzzy clustering
quantification