改进的基于直方图和空间特征的图像检索

Image Retrieval Based on Improvement Histogram and Spatial Feature

在主要面积直方图算法的基础上,分析了图像灰度连通区域的空间分布特征,并结合主要面积直方图进行图像检索,进一步提高了算法的检索性能。同时,结合人眼视觉特性,对图像的灰度进行了量化,在不降低检索性能的基础上降低了计算和时间复杂度。

A new image retrieval algorithm based on the main region＇s area histogram was proposed. In order to reduce the calculation and time complexity, the gray is quantized according to human vision firstly. Then, the spatial distribution of the adjacent regions with the same gray level is further analyzed and extracted as the spatial feature. Finally, the spatial feature and its main region＇s area are used as indexes. Experiment results show that this method is more effective in image retrieval than histogram and the main region＇s area histogram.