基于JPEG2000的图象局部可分级检索

JPEG2000-Based Scalable Lookup of Image's Local Regions

提出了一种基于 JPEG2 0 0 0的图象局部可分级检索方法 .首先 ,对图象采用 JPEG2 0 0 0的基本算法进行压缩编码 ;然后 ,把压缩码流按零树结构编排 ,每个零树结构对应着图象的某个网格区域 ,从零树结构所包含的 DWT系数可以分级重建这个网格区域的内容 .检索时 ,先从图象的最低分辨率子图象 (即 L L 子图象 )的预览开始 ,一旦发现目标 ,则锁定目标 ;然后在空间分辨率和信噪比两个方面逐级浮现目标所在网格的内容 .由于该方法仅仅使用部分压缩码流数据来恢复图象的局部内容 ,从而节省了大量的计算资源 ;最后给出了基于 JPEG2 0 0

A method of JPEG2000 based scalable lookup of image's local regions is presented in this paper. At first, we compress the image with the baseline algorithm of JPEG2000. Then we arrange the compressed data in light of zerotree structures, each zerotree is corresponding to a local region(mesh) of original image. When we want to look up a target on the image, we first browse the lowest resolution subimage(LL subimage) and locate mesh which the target belongs to. Then we reconstruct the mesh level by level in both spatial revolution and Signal to Noise (SNR). In this way we use only part of the compressed data to reconstruct the local region of image and show the target. It can save much computation resources. In the end of this paper, the experimental results of JPEG2000 based scalable lookup of image local regions are given.