摘要
对随机分形压缩算法中存在较严重的方块效应及原因做了分析,并由此提出了改进算法。按定义域块和值域块的方差,将其分为平坦、边缘和中间3类,对每一值域块,先搜索属于该值域块邻域且为同一类的定义域块,如满足阈值条件,则将其记录下来,如不满足,则按随机分形压缩算法寻找其它定义域块。实验表明,提出的算法能在保持恢复图像质量的同时缩短压缩时间,并减少方块效应。
The serious square-effect and reason in the randomized fractal compression algorithm is analyzed and the improved algorithm is presented. The domain and range blocks are partitioned into three groups according to their variance： Shade, edge and midrange. For each range block, it first searched the domain block that is adjacent to the range block and same type with the range block. The method recorded it while the threshold condition is satisfied. If not satisfied finding other domain blocks by using randomized fraetal compression algorithm. The experiments show that the encoding time while retaining the quality of reconstruction image is reduced and the square-effect is decreased.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第7期1586-1588,共3页
Computer Engineering and Design
关键词
压缩
随机化
分类
邻近块搜索
分形
compression
randomized
classified
search adjacentblocks
fractal
