In this paper, a fractal image compression system based on adaptive variableshapeblock transform is proposed. The procedure can be divided into approximately two basic phases: the classification phase and the adapti...
详细信息
In this paper, a fractal image compression system based on adaptive variableshapeblock transform is proposed. The procedure can be divided into approximately two basic phases: the classification phase and the adaptive shapeblock phase. In the classification phase, a testing image block is assigned to its own class according to its direction attribute. The direction of the block is estimated by employing the technique of projection/backprojection which is a tomographic image reconstruction skill. Due to computational complexity, only four projection directions are resolved to construct the projection classification algorithm (PCA). Computer simulations show that the proposed classification method outperforms the gradient classification algorithm. In the adaptive shapeblock phase, PCA and the algorithm of split/merge are incorporated to form the variableshapeblock transformation. For the range blocks and domain blocks, we define fifteen merging patterns and ten shape components in the searching algorithm. Under this fractalcoding approach, a better trade-off between the decoded image quality and transmitting bit rate is performed. Several examples are presented to show the effectiveness of this coding system.
暂无评论