摘要
本文提出了一种改进的颜色直方图图像检索方法。通过对HSV颜色空间的变换,把颜色空间量化为9个区间,在此基础上计算出每种颜色区间的空间分布信息。综合颜色直方图和空间信息作为图像相似性的量度。试验表明,该算法提取的特征数较少,检索速度快,并且准确度较高。
An Image Retrieval Method Based on Color Histogram and Space Information is proposed in this paper. Color space is quantized into 9 intervals by transforming HSV color space. Then it gets the space information of each interval. It uses color histogram and space information as image similarity measure. Experiments show that fewer features are extracted by this method, and the retrieval speed and accuracy is improved.
出处
《微计算机信息》
北大核心
2008年第4期246-247,共2页
Control & Automation
基金
国家科技支撑计划支持(2006BAH02A13
2006BAH02A03)