摘要
提出了一种基于方块编码的图像检索算法。首先将图像分成互不重叠的子图像块,根据图像块中各像素间的色差,利用方块编码的思想对这些子图像进行编码,然后根据人眼的视觉特性来定义图像的关键块,最后借助于基于关键字的文本检索技术进行图像检索。同时,考虑到不同类型的关键块在表征图像内容时重要程度的不同而赋予其不同的权值。实验结果表明本文算法在图像的相似性检索时是有效的,并具有较高的检索效率。
A novel image retrieval method based on Block Truncation Coding (BTC) was proposed. Firstly, the image was divided into non-overlapped and equally-sized blocks, and these blocks were coded by the method of BTC. Then, the keyblocks of image were defined according to the human visual feature and the color difference between pixels in the image blocks. At the same time, the different importance of each keyblock in describing the image was taken into account, and the weighted function was introduced. Finally, the technologies in Text Information Retrieval (TIR) were applied in Content-based Image Retrieval (CBIR). Experimental results show that the proposed method has reasonable and robust retrieval performance and is more effective in the image retrieval than the other algorithms discussed in the paper.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2007年第1期117-120,共4页
Opto-Electronic Engineering
关键词
图像检索
方块编码
关键块
特征提取
Image retrieval
Block truncation coding (BTC)
Keyblock
Feature extraction