一种基于曲波变换的图像去块算法  

An image de-blocking algorithm based on curvelet transformation

作　　者：邵真天[1] 袁杰[1] 

机构地区：[1]南京大学电子科学与工程学院,江苏南京210093

出　　处：《智能系统学报》2012年第2期102-107,共6页CAAI Transactions on Intelligent Systems

基　　金：江苏省自然科学基金资助项目(BK2010386)

摘　　要：图像块效应是由于进行图像压缩编码时采用离散余弦变换(discrete cosine transform,DCT)并对其系数进行量化处理而引起的,该过程丢失了一些频率成分,并引起了子块边界不连续的跳变.针对这一问题,提出了一种基于曲波变换的图像去块算法,该算法首先对退化图像进行曲波变换,再对所获取的各层曲波系数进行处理以备复原图像.通过寻找各层中与原始图像块效应相对应的系数,对不同的层采用不同的方法,并计算图像重建时所要使用的新系数矩阵.实验表明,该算法比传统在客观和主观评估中都被普遍运用的空间域和小波去块方法,得到了更佳的图像复原效果,且保留了更多的细节.The image block effect is caused by quantification processing of coefficients when using discrete cosine transformation （DCT） to compress coding of images. DCT dumps some frequencies, leading to discontinuous leaps of sub-block boundaries. An image de-blocking algorithm based on eurvelet transformation was proposed. This algorithm first carried out curvelet transformation of degraded images, then processed the obtained curvelet coefficients separately layer by layer, so as to recover the images. The coefficients corresponding to the block effect of the original image could be found for every layer, proving that different methods apply to different layers. Then new coefficient matrixes were calculated, using the images which were reconstructed. Experiments show that the algorithm retains more details and generates better recovery results than the traditional spatial domain and wavelet de-blocking methods which are used commonly for both subjective and objective evaluations.

关 键 词：去块 曲波变换 图像复原 多尺度分析 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]