拼贴误差是值域块与“最匹配的”定义域块相似性的一个度量,然而很多基于分形编码参数的分形图像检索技术却忽略了拼贴误差。该文使用拼贴误差直方图作为检索条件,来判定图像的相似度。对有200多张纹理图像的数据库的实验显示,该方法不仅减少了计算的复杂度,而且也有一定的检索准确率。
Collage error is a quantitative measure or'the similarity between range block and "best-match" domain block. However, it is neglected by many fractal-based image retrieval techniques. In this paper, a new fractal coding based indexing technique using histogram of collage errors is proposed. Experimental results on a database of more than 200 texture images show that the proposed method not only reduces computational complexities, but also enhances the retrieval rate.