【摘要】 特征提取是基于内容图像检索(CBIR,Content-BasedImageRetrieval)系统实现的关键技术之一,同时更加准确有效的图像检索需要结合图像的各类特征。论文提出在重新量化彩色图像的基础上提取多类特征进行检索。首先给出颜色空间的选择和重量化方案;其次对重量化结果进行分析以确定多类特征;然后提出图像特征间的相似度计算公式;最后对检索过程进行仿真实验。更多还原

【Abstract】 Feature extraction is a key technology in CBIR.An effective and efficient retrieval system needs to combine several kinds of image features.This paper proposes a method that uses the result of color re-quantization to extract several kinds of features for retrieval.First,it presents the selection of color space and an approach of re-quantization. Second,it analyzes the result of re-quantization to extract multi-feature from color image.Next,it gives the similarity evaluation to feature matching.Finally,authors do a simulation experiment.更多还原